DW Weekly #172 – 9 August 2024

 Page, Text

Dear readers, 

Welcome to another issue of the Digital Watch weekly! 

Remember that ScarJo vs OpenAI kerfuffle over her voice being used for ChatGPT? Here’s a quick summary: OpenAI used a voice eerily similar to actress Scarlett Johanson’s voice, seemingly inspired by her performance as an AI assistant in the movie Her. The catch is that OpenAI did it without her consent. The company has since removed the voice from ChatGPT. Most recently, Johannson said OpenAI CEO Sam Altman would make a good Marvel villain.

This week, The New York Times reported that Meta is negotiating with actors such as Judi Dench, Awkwafina, comedian Keegan-Michael Key, and other celebrities and influencers to use their voices for its MetaAI digital assistant. Reportedly, Meta could pay millions in fees to secure their voices.

Our voices are one of the very few things that are uniquely ours, and as such, are considered biometric data. We are certainly more accustomed to other data being used for biometric identification – chief among them our fingerprints and faces to unlock our sleeping devices. Voice identification is plagued with problems that make it less accurate than other kinds of biometric identification. For instance, if we contract a cold, it can be enough for the software not to recognise the voice. Still, voice identification is a viable identification method.

The Little Mermaid gave up her voice in a deal with the evil sea witch – she did it for love. Celebrities might sell their voice for money. Readers, would you do it? Would you feel comfortable with your voice saying thoughts that are not yours?

A corporation is depicted as an evil, black sea witch offering a contract to a mermaid who is holding bubbles containing a heart and musical notes.

The UN Ad Hoc Committee (AHC) adopted a draft convention on cybercrime after extensive negotiations – keep an eye out for a short update on the outcomes next week and a more extensive analysis in the next few weeks.

If you’re an AI professional or enthusiast watching the Paris Olympics or an Olympian reading this digest, scroll down to our reading corner or click here to read our take on the transformative impact of AI on the Olympics.

Andrijana and the Digital Watch team


Highlights from the week of 2-9 August 2024

google mobile application smart phone screen

Google controls about 90% of the online search market and 95% of smartphones.

apple inc 508812 1280

Apple’s lucrative $20 billion annual deal with Google, which sets Google’s search engine as the default on Apple devices, could be under threat after a US judge ruled that Google…

CrowdStrike Logo 2023 Primary Black

The global computer outage caused by CrowdStrike’s software update led to lawsuits from air travellers.

dfb20d24 1c75 477d 9813 3f7d2e4db8c5

The first patient, Noland Arbaugh, can now play video games and browse the internet using the implant.

Altman 1

Project Strawberry includes a ‘deep-research’ dataset, expected to revolutionise AI research capabilities.

chinese flag with the city of shanghai in the background and digital letters ai somewhere over the flag

The growing number of approved algorithms highlights an efficient approval process and China’s striving for leadership in AI regulation.

b27c64ea 20bc 4ceb b68d 05b19b266ce0

The revived suit alleges that Altman has altered the narrative to capitalise on the company’s technological advancements, moving away from the altruistic mission it was founded upon.

tiktok i like the first picture can you somewhere to say tiktok

They argue that TikTok poses a national security threat due to its ties to China, which could exploit user data.

asian woman using tiktok application her iphone14 while relaxing coffee shop

The programme allowed users to earn points for activities such as watching videos and liking content.

national flag of the philippines

New digital bank licenses in the Philippines will require innovative business models and a focus on reaching underserved markets.



ICYMI

Discover how John Calvin’s 16th-century ideas on individualism and societal responsibility journeyed from Geneva to Silicon Valley and shaped today’s tech-driven world!

What exactly is a deepfake and how do you spot one? Find out in the 7th part of our AI shorts series.


Survey

Olimpic AI Agenda

Discover how AI is transforming every facet of the Olympics, from organisation, logistics, data collection, and event broadcasting, to the sustainability of the games. Join us for an in-depth exploration of the Olympic AI agenda.

fact fake wordplay banner headline

Content governance and disinformation came into the sharp public focus as hate speech spilt over from social media platforms to the streets of UK cities. Marilia Maciel provides just-in-time reflections on what can be done to improve online content governance and reviews key findings from Diplo’s research on online disinformation and content management. Read the blog and register for debates in September.


Andrijana20picture
Andrijana Gavrilovic – Author
Editor – Digital Watch; Head of Diplomatic u0026amp; Policy Reporting, DiploFoundation
nPHsW2zD ginger
Virginia Paque – Editor
Senior Editor Digital Policy, DiploFoundation

DW Weekly #171 – 2 August 2024

 Page, Text

Dear readers, 

Welcome to another issue of the Digital Watch weekly! 

The co-conveners of the World Trade Organization (WTO) Joint Initiative (JI) on Electronic commerce – Australia, Japan and Singapore – have published a stabilised text of an Agreement on Electronic Commerce, a significant milestone after almost seven years of discussions and negotiations.

Either you know exactly what we’re writing about, or it is all gobbledygook to you. Let’s break it down.

What’s a WTO JI? WTO Joint Initiatives (JIs) are a way for a group of World Trade Organization members to move forward on specific issues without waiting for the entire organisation to reach a consensus. They are open to any WTO Member. 

What does the WTO JI on Electronic commerce negotiate? This JI tackles a mix of traditional trade issues and modern digital policy challenges. Discussions cover trade facilitation, cross-border data flows, data localisation, access to source code, and net neutrality.

Why is the WTO JI particularly important? So far, e-commerce and digital trade regulations have been handled mainly through preferential trade agreements (PTAs) between countries. However, there isn’t a specific WTO agreement on e-commerce. Creating one would help standardise e-commerce rules globally, making it easier for everyone to do business in the digital age.

What’s in the Agreement on Electronic Commerce? The text contains provisions to:

  1. Promote easy digital trade within and between countries, including electronic signatures and invoices.
  2. Make international digital trade more reliable and affordable by working together on cybersecurity risks.
  3. Permanently ban customs duties on digital content among participating countries.
  4. Protect online consumers from misleading and fraudulent activities.
  5. Protect the personal data of workers and consumers
  6. Help consumers and companies from developing countries participate in digital trade.
  7. Digitalise the various trading systems to make global trade cheaper, faster, and more secure for businesses.
  8. Encourage competition in the telecommunications sector by ensuring independent regulators, better access to infrastructure, and market-based frequency band assignments.

What’s missing in the text? Negotiations on crucial digital issues like data flows and source code hit a roadblock when the USA pulled its support to maintain domestic policy flexibility. While data flows are essential for the digital economy, global rules on this matter are unlikely to be agreed upon anytime soon. The co-conveners simply state that ‘participants recognise that some issues of importance to digital trade have not been addressed in this text. Participants will discuss the inclusion of these issues in future negotiations.’

Who’s missing from the deal? The latest draft text, published by the co-conveners, represents 82 out of 91 JI members. However,  Brazil, Colombia, El Salvador, Guatemala, Indonesia, Paraguay, the Separate Customs Territory of Taiwan, Penghu, Kinmen and Matsu, Türkiye, and the USA are still reviewing the text domestically.

Under the title 'Agreement on Electronic Commerce, delegates discuss around a table with a computer and papers. Chat bubbles contain icons of agreement, shopping carts, a graph, a fingerprint a piggy bank, and an umbrella.

What are the next steps? To become the basis for global rules on digital trade among WTO members, the text must be integrated into the WTO legal framework. However, the JI on e-commerce runs into a snag here.

The JIs themselves run into opposition from a number of WTO members who hold that JSIs do not have any legal status because they were not launched based on consensus. Similarly, these countries claim that the outcomes of JIs are not based on consensus and are neither multilateral agreements nor plurilateral agreements as defined in Article IV of the agreement that established the WTO – the Marrakesh Agreement.

If the agreement crosses the finishing line, our colleague Marilia Maciel writes that five changes in the global landscape are important to consider in its implementation. These are rising digital inequality, a shift eastwards in digital trade rule-making, the growing importance of Digital Economy Agreements (DEAs), the uncertain consequences of the re-wiring of Global Value Chains, and the systemic nature of challenges to multilateralism.

In other news, researchers uncovered the largest ever ransomware payment of a whopping USD 75m, the EU AI Act officially comes into force on 1 August, and the concluding session of the Ad Hoc Committee on Cybercrime has reached its halfway point today.

Andrijana and the Digital Watch team


Highlights from the week of 26 July-2 August 2024

hacker 8003399 1280

Researchers also note that the USA remains the top target for ransomware, followed by the UK and Germany.

8U8BWHpU Meta

Filed in 2022 under Texas’ 2009 biometric privacy law, the lawsuit accused Meta of capturing biometric data through the ‘Tag Suggestions’ feature, which has since been discontinued.

global connections business internet country united kingdom united kingdom map stockpack istock scaled

The scrutiny is part of a global regulatory effort, highlighted by a joint statement from the CMA, US, and the EU regulators to safeguard fair competition in the AI industry.

building 1011876 1280

Historical tech partnerships often evolve into competitive relationships.

the us capital building at night with a full moon politics senate house of representatives

Supporters, including some tech executives and advocacy groups, see the legislation as crucial for protecting children online.

malta osce

Minister for Foreign and European Affairs and Trade from Malta, Ian Borg, advocated for increased cyber resilience among OSCE member countries through enhanced cooperation.

AIatUN 1

Stakeholders, including the UN High Commissioner for Human Rights and the Electronic Frontier Foundation, stress the need for more transparency and robust protections in the treaty.

cybersecurity cybercrime internet scam anonymous hacker crypto currency investment digital network vpn technology computer virus attack risk protection

Despite the breach, the ICO found no evidence of data misuse. In response, the Electoral Commission has since upgraded its security infrastructure.

north korea flag is depicted on the screen with the program code

The hackers also infiltrated NASA’s systems, extracting over 17 gigabytes of data.

Biden signs 2023 Executive order on AI CNBC

Key industry figures highlight her stable leadership and alignment with Silicon Valley values.



Reading corner

selective focus of toy shopping cart with small carton boxes near laptop e commerce concept

The WTO Joint Initiative stabilised ‘Agreement on Electronic Commerce’ must contend with five changes that took place in the global landscape that are important to consider in its implementation.

robot hands fingers point laptop button advisor chatbot robotic artificial intelligence concept

In his second chat with AI, Petru Dimitiriu delves into a conversation revealing AI assistants’ strengths and weaknesses.


Survey

📢 We want your feedback!

Are you following the UN OEWG on cybersecurity discussions? We want to hear from you! 

Help us improve our reporting by participating in our short survey. Your insights are invaluable and will directly contribute to making our coverage even better. 

Thank you for your time and support!


Andrijana20picture
Andrijana Gavrilovic – Author
Editor – Digital Watch; Head of Diplomatic & Policy Reporting, DiploFoundation
nPHsW2zD ginger
Virginia Paque – Editor
Senior Editor Digital Policy, DiploFoundation

DW Weekly #170 – 26 July 2024

 Page, Text

Dear readers, 

Welcome to another issue of the Digital Watch weekly! 

It should have been a routine update. Alas, an update Crowdstrike issued to its Falcon Sensor kernel-level driver that protects Windows computers caused a major tech outage last Friday, disrupting operations across different industries worldwide, including air travel, healthcare, financial services, and media. 

Caricature drawing of three human figures studying CrowdStrike computer screens and a black world analysis map. Pin markers and lines connect a global network with interspersed Microsoft blue screens of death.

The problem began when CrowdStrike released a content configuration update for the Windows sensor to collect telemetry on potential new threat techniques. These updates are a regular aspect of the Falcon platform’s dynamic protection system, sometimes occurring daily. However, this particular update contained a logic error that resulted in a system crash, causing the notorious blue screen of death (BSOD) on 85 million Windows devices.

The issue was traced back to a bug in CrowdStrike’s Content Validator, which allowed the problematic update to pass validation despite containing problematic content data. 

The silver lining? The outage, while severe, did not lead to exploitation by malicious actors on a larger scale. However, CrowdStrike intelligence reported observing threat actors leveraging the situation in Latin America, underlining the vulnerability of digital systems to exploitation.

Financially, the impact of this incident is staggering. According to cyber insurer Parametrix, the faulty update could result in losses of up to $54 billion for companies like Microsoft, major airlines, banks, and healthcare providers. In a bizarre twist, CrowdStrike tried to make amends by sending $10 Uber Eats gift cards to affected partners. However, many recipients got error messages saying the vouchers were cancelled, with Uber flagging the high redemption rate as fraud.

Screenshot of a CrowdStrike memo thanking clients with the offer of a USD 10 UberEats credit.

What did we learn from this case? Numerous organisations are overreliant on single-point IT solutions. Should tech companies bear responsibility for cyber risks associated with their products and services? 

Defining supplier responsibility (and liability) for the security and stability of digital products through legal instruments is essential for ensuring accountability and safety. This would incentivise companies to invest more in robust security measures, thorough testing protocols, and fail-safe mechanisms. Moreover, it would foster a culture of responsibility within the tech industry, where the potential real-world impacts of digital failures, their solutions, and their broader implications for society are given the serious consideration they deserve. 

As these debates heat up, the scrutiny of cybersecurity practices is intensifying:  CrowdStrike’s CEO has been summoned by a US congressional committee to explain the mess, underscoring the growing legislative interest in the issue. 

In other news, France is bracing for cyberattacks as the Olympics kick off. The USA, the EU, and the UK authorities pledged to protect competition in generative AI foundation models and AI products. Meanwhile, Sam Altman is advocating for a USA-led global coalition on AI to ensure a democratic vision for AI prevails over an authoritarian one.

Andrijana and the Digital Watch team


Highlights from the week of 19-26 July 2024

eu ai act

Details on how the working group will function will be finalised after the summer.

3d render artificial intelligence logo deep learning blockchain neural network concept generative ai

The statement recognises the profound potential of AI and foundation models, pinpointing several competition risks: concentrated control of key AI resources, market power entrenchment by incumbent digital giants, and potentially anti-competitive collaborations among major AI players.

Sam Altman WEF

Altman also advocates for international collaboration and possibly establishing governance bodies to ensure AI benefits are widely shared and risks minimised.

nvidia chip

Despite initial weak sales, Nvidia is now on track to sell over 1 million H20 chips in China this year, generating more than $12 billion.

search bar google chatgpt competitors

SearchGPT offers summarised results with source links and follow-up queries.

WTO building

The Joint Statement Initiative on Electronic Commerce, co-convened by Australia, Japan, and Singapore, announced a milestone on 26 July 2024, with participants reaching a stabilised text for the Agreement on Electronic Commerce after five years of negotiations. The negotiations will continue.

singapore

Expected to lower business costs and boost services trade, it underscores Singapore’s significant role as the EU’s fifth-largest services trade partner, with over half of the €43 billion services traded in 2022 delivered digitally.

app store 1174440 1280

The probe will examine whether Apple imposed unequal commercial conditions on developers of mobile applications sold through its App Store.

european union and meta platforms

The potential fine for Meta could reach $13.4 billion, or 10% of 2023 global revenue.

nigerian flag laptop keyboard online business education shopping nigeria concept 3d rendering

The fine, one of the largest by an African regulator on a global tech company, follows a 38-month investigation revealing Meta’s non-compliance with Nigeria’s Data Protection Regulation (NDPR).

cyber attack anonymous people hacking security protection hackers access sensitive data cyber war

The breach is suspected to stem from an earlier hack of a system by third-party vendor Diligent Corp.

paris flag and arc de triomphe

The country is committed to minimising their impact, says Prime Minister Gabriel Attal.

landscape 5311428 1280

Battling progressive supranuclear palsy (PSP), a condition that limits her ability to speak and move, Wexton announced in September that she will not seek re-election due to her worsening health.



ICYMI

Cyber tools for disaster management july 2024 1920x1080px intro slide

This webinar explored how advanced technology can improve disaster preparedness, response, and recovery in Africa, highlighting innovative cyber tools and solutions that enhance crisis management and resilience.

DW Weekly #169 – 19 July 2024

 Page, Text

Dear readers, 

Welcome to another issue of the Digital Watch weekly! 

The 8th substantive session of the Open-ended Working Group (OEWG) on security of and in the use of information and ICTs 2021–2025 ended last Friday. The group gathered to adopt its annual progress report (APR), which takes stock of the group’s discussions over the previous year but also charts the way forward for the group’s discussions for the following year.

The OEWG is a process encompassing all interested UN members, and it adopts its decisions by consensus, so adopting APRs is always a momentous effort. This time, 80 UN member states participated, with their views on what the APR should or should not contain. The chair opened the session by pleading for members to maintain a dynamic of mutually agreed additions to the text of the APR, not mutually decided deletions leading to the document being empty.

Things were tense through the penultimate meeting (the OEWG sessions are typically made of ten meetings), while Iran held that they could not accept the document. The crux of Iran’s unwillingness was paragraph 27, which Iran feels does not accurately reflect the international legal order because the text says that states have obligations stemming from voluntary norms. However, at the very last meeting, Iran distanced itself from paragraph 27, and the APR was adopted.

Drawing shows standing observers studying a bar chart on the wall. The chart shows the relationship of threats, norms, international law, confidence-building measures, capacity building and regular institutional dialogue from 2019 through 2024 with a red bar. A UN logo is shown on the wall to the left of the chart.

What is not solved in the OEWG will undoubtedly make its way into the agenda of the next mechanism, but not everything should be deferred, or the delegations will be in for quite a headache in 2026. Town halls and intersessional meetings can be held to discuss international law and to agree on the elements of regular institutional dialogue, and we know Chair Burhan Gafoor is good at that – last year, the pace of meetings was near-breakneck speed.

A summary of the discussions is available on Digital Watch. AI reports and transcripts from the 8th substantive session are available on our dedicated web page – a more in-depth human-generated analysis is planned for next week.

UNFOLDING NOW
A major tech outage today (19 July) is disrupting operations across different industries worldwide, including air travel, healthcare, financial services, and media. The outage’s root cause was traced to a software update by cybersecurity firm CrowdStrike, whose CEO stated, ‘The issue has been identified, isolated, and a fix has been deployed.’ We’ll follow the story as it develops.

Andrijana and the Digital Watch team


Highlights from the week of 12-19 July 2024

windows blue screen

Key sectors including media, healthcare, and financial services also experienced significant impacts.

OpenAi and Microsoft

As OpenAI’s Strawberry advances the reasoning capabilities of models the company must face off with its employees who consider such models a threat to humanity.

Screenshot 2023 08 09 at 12.57.10

Nvidia could face charges if the investigation yields sufficient evidence, hinting at the gravity of the situation.

words 2904676 1280

The incident has led to a surge in wild claims and disinformation online

trump 2704264 1280

The FBI was assisted by the Israeli firm Cellebrite.

businessman protecting data personal information cyber security data concept padlock internet te 1

Kaspersky will continue operations outside the US, focusing on other strategic markets.

tiktok icon coming out screen mobile phone 3d illustration

TikTok’s owner, ByteDance, must comply with EU gatekeeper regulations CJEU rules.

security data technology identity verification personal privacy concept shield icon fingerprint scan symbol random numbers code virtual screen with man pointing finger laptop computer

US senators press AT&T, Snowflake on data breach, fearing sensitive customer data auctioned or compromised, demand accountability.

nokia telecom egypt

Nokia and Telecom Egypt have partnered to introduce 5G technology to Egypt, enhancing connectivity in major cities with Nokia’s AirScale portfolio. This collaboration aims to revolutionise Egypt’s digital landscape and support innovation.

european union table flag

The EU’s new regulation requires Digital Product Passports for most products by 2030.

kenza layali

Kenza Layli, an AI-generated Moroccan influencer, has been crowned the first Miss AI. Created by Myriam Bessa, Layli promotes diversity and inclusivity in the AI creator landscape, outshining 1,500 entries worldwide.



ICYMI

Do we need a new international treaty to protect critical infrastructure (CI)? Which challenges would this help to address, and why? What could be the scope of this treaty? We asked our guests, Aude Géry, Senior Researcher in Public International Law and Digital Security, Geopolitics of the Datasphere (GEODE), and Marc Henauer, Senior Politics and International Affairs Officer, Swiss National Cyber Security Centre (NCSC).


Upcoming

Cyber tools for disaster management july 2024 916x530px 1

This webinar will focus on leveraging technology to enhance preparedness, response, and recovery efforts in the face of natural or man-made disasters on the African continent. This webinar aims to explore the use of cutting-edge cyber tools and technologies to improve disaster management in Africa. It will delve into innovative solutions that enable quicker, more coordinated responses to crises, ultimately enhancing overall resilience in the region.

Numéro 91 de la lettre d’information Digital Watch – juillet 2024

 Advertisement, Poster, Page, Text, Person, Face, Head

Coup d’œil

Coup d’œil : les développements qui font des vagues

Gouvernance de l’IA

Sous l’égide de la Chine, l’Assemblée générale des Nations unies a adopté une résolution non contraignante sur le renforcement des capacités d’IA afin d’améliorer celles des pays en développement grâce à la coopération internationale. Elle appelle également au soutien des organisations internationales et des institutions financières. Les ministres africains des TIC et des Communications ont approuvé la Continental AI Strategy et le pacte numérique africain afin de stimuler la transformation numérique du continent. Le communiqué des dirigeants du G7 a mis l’accent sur une stratégie coordonnée pour gérer les opportunités et les défis de l’IA, en introduisant un plan d’action pour l’adoption de l’IA sur le lieu de travail et en soulignant des initiatives telles que l’avancement du code de conduite international du processus d’Hiroshima, le soutien aux PME, et la promotion de l’inclusion numérique et de l’éducation et de la formation tout au long de la vie.

Le Fonds monétaire international a recommandé des politiques fiscales aux gouvernements confrontés à l’impact économique de l’IA, notamment des taxes sur les bénéfices excédentaires et une taxe sur le carbone.

La Chine est en tête du classement mondial des demandes de brevets d’IA générative, dépassant largement les États-Unis. Dans le même temps, les entreprises technologiques américaines dominent la production de systèmes d’IA de pointe, selon l’Organisation mondiale de la propriété intellectuelle (OMPI). Un rapport de la Commission européenne montre que l’UE est à la traîne par rapport à ses objectifs 2030 en matière d’IA, avec seulement 11 % des entreprises utilisant des technologies d’IA désignées, loin de l’objectif de 75 %. Le ministère japonais de la Défense a introduit sa première politique en matière d’IA afin d’améliorer les opérations de défense. Le Brésil s’associe à OpenAI pour moderniser les processus juridiques, réduire les frais de justice et améliorer l’efficacité du bureau du procureur général.

Technologies 

Les États-Unis ont présenté un projet de règles visant à réglementer les investissements en Chine, en se concentrant sur les secteurs de l’IA et des technologies de pointe susceptibles de poser des menaces à la sécurité nationale. Les États-Unis prévoient d’étendre les sanctions sur les puces semi-conductrices et d’autres biens vendus à la Russie, en ciblant les vendeurs tiers chinois. Des discussions sont en cours avec les Pays-Bas et le Japon pour restreindre 11 usines chinoises de fabrication de puces et étendre les contrôles à l’exportation d’équipements. Les États-Unis sont confrontés à une pénurie de 90 000 techniciens en semi-conducteurs d’ici à 2030, ce qui a incité l’Administration Biden à lancer un programme de développement de la main-d’œuvre.

La Commission européenne prend connaissance des points de vue de l’industrie sur l’augmentation de la production chinoise de puces informatiques d’ancienne génération.

La Chine va élaborer des normes pour les interfaces neuronales directes (BCI) par l’intermédiaire d’un nouveau comité technique, en se concentrant sur le codage des données, la communication, la visualisation, la collecte de données d’électroencéphalogramme et les applications dans divers domaines.

Infrastructure

Des entreprises de télécommunications du Kazakhstan et de l’Azerbaïdjan vont investir plus de 50 millions d’USD dans la pose de 370 kilomètres de câbles à fibres optiques sous la mer Caspienne. Dans le même temps, le nouveau responsable de l’économie numérique du Sénégal a annoncé son intention de renforcer l’infrastructure numérique, de coordonner les programmes gouvernementaux, d’encourager les collaborations et de s’appuyer sur les réalisations antérieures pour accroître la contribution de l’économie numérique au PIB.

Cybersécurité

Le Conseil de sécurité des Nations unies a tenu un débat ouvert sur la cybersécurité, axé sur l’évolution des cybermenaces et la nécessité de réaliser des avancées numériques positives.

La récente cyberattaque contre la société de stockage en nuage Snowflake s’annonce comme l’une des plus importantes violations de données jamais survenues, touchant des centaines de leurs entreprises clientes et des millions d’utilisateurs individuels. Le centre national de données de l’Indonésie a été touché par une variante du rançongiciel LockBit 3.0, ce qui a perturbé les contrôles d’immigration et les services publics. Les pirates se sont depuis excusés et ont proposé de divulguer les clés des données volées. Le centre hospitalier universitaire de Zagreb, en Croatie, a également été victime d’une cyberattaque par LockBit. Malgré la hausse des attaques de rançongiciels, un rapport de Howden indique que les primes d’assurance cybernétique mondiales sont en baisse, car les entreprises améliorent leurs capacités d’atténuation des pertes. En outre, près de dix milliards de mots de passe uniques ont été divulgués dans une collection appelée RockYou2024, ce qui accroît les risques pour les utilisateurs qui les réutilisent.

L’Australie a demandé aux sociétés Internet de créer des codes applicables dans les six mois pour empêcher les enfants d’accéder à des contenus inappropriés. La Nouvelle-Zélande a transformé l’appel à l’action de Christchurch contre les contenus terroristes en ligne à une ONG, désormais financée par des entreprises technologiques telles que Meta et Microsoft.

Droits numériques

La proposition de loi de l’UE visant à rendre obligatoire l’analyse par l’IA du contenu des applications de messagerie afin de détecter les contenus pédopornographiques fait l’objet de critiques en raison des risques d’atteinte à la vie privée et des faux positifs potentiels. Les régulateurs de l’UE ont accusé Meta d’avoir enfreint les règles en matière de technologie par le biais d’un modèle publicitaire « pay or consent » sur Facebook et Instagram, alléguant qu’il forçait les utilisateurs à consentir au suivi des données. Le ministère américain de la Justice envisage de porter plainte contre TikTok pour violation présumée de la vie privée des enfants. Google est accusé par le groupe européen de défense des données NOYB (none of your business) de suivre les utilisateurs sans leur consentement éclairé par l’intermédiaire de sa fonction Privacy Sandbox (bac à sable de protection de la vie privée).

Juridique

La Cour pénale internationale enquête sur des cyberattaques russes présumées contre des infrastructures ukrainiennes, qu’elle considère comme des crimes de guerre potentiels. En Australie, une action en justice a été engagée contre Medibank pour une violation de données touchant 9,7 millions de personnes. ByteDance et TikTok contestent une loi américaine visant à interdire l’application, en invoquant des préoccupations liées à la liberté d’expression. Les sociétés mondiales de diffusion en continu contestent la nouvelle réglementation canadienne, exigeant que 5 % des revenus soient consacrés aux informations locales, remettant en cause la légalité des actions du gouvernement.

Économie de l’internet

Le ministère chinois du Commerce a présenté un projet de règles visant à soutenir le commerce électronique transfrontalier en encourageant la création d’entrepôts à l’étranger, et en améliorant la gestion des données et la supervision des exportations. Nvidia fait face à des accusations potentielles en France concernant des allégations de comportement anticoncurrentiel. Le premier semestre 2024 a été marqué par une forte augmentation des vols de cryptomonnaies, avec plus de 1,38 milliard de dollars volés au 24 juin.

Développement

La première partie du rapport annuel de la Broadband Commission sur l’état de la large bande, intitulé « Leveraging AI for Universal Connectivity », explore l’impact de l’IA sur l’administration en ligne, l’éducation, les soins de santé, la finance et la gestion de l’environnement, ainsi que son potentiel pour combler ou élargir la fracture numérique. L’Inde imposera le port USB-C comme port de charge standard pour les smartphones et les tablettes à partir de juin 2025, s’alignant ainsi sur les efforts de l’UE pour réduire les déchets électroniques.

Socioculturel 

Les législateurs de l’État de New York ont adopté une loi interdisant aux plateformes de médias sociaux d’afficher des contenus algorithmiques addictifs à des utilisateurs de moins de 18 ans sans le consentement de leurs parents. La Commission européenne a demandé à Amazon des détails sur la manière dont elle se conforme aux règles de la loi sur les services numériques, en mettant l’accent sur la transparence de ses systèmes de recommandation. Google Translate ajoute 110 langues, grâce aux progrès de l’IA.

LES CONVERSATIONS DE LA VILLE – GENÈVE

Du 4 au 14 juin, le Conseil de l’Union internationale des télécommunications (UIT) a pris des décisions importantes sur le développement spatial, l’action numérique verte et la coopération numérique mondiale. Il a examiné le rapport de la Secrétaire général de l’UIT sur la mise en œuvre du Programme Espace 2030, en mettant l’accent sur l’exploitation des technologies spatiales au service de la durabilité. Des résolutions ont été rédigées pour souligner le rôle de l’UIT dans l’utilisation des technologies numériques pour la durabilité, avec un rapport sur les initiatives numériques vertes actuelles. L’UIT continuera de participer au Pacte numérique mondial (PMN) afin de renforcer la coopération mondiale dans ce domaine.Le 14 juin, la première Journée des mondes virtuels de l’ONU a présenté des technologies telles que la réalité virtuelle et augmentée, le métavers et l’informatique spatiale pour faire progresser les ODD. L’événement comprenait un segment de haut niveau, des applications réelles, des discussions sur la politique et le lancement de l’Initiative mondiale sur les mondes virtuels – Découvrir le CitiVerse, une plateforme pour développer des cadres, sensibiliser, partager les meilleures pratiques et tester des solutions metaverses dans les villes.

En bref

IA@ONU : naviguer sur la corde raide entre innovation et impartialité

Les Nations unies ne manquent pas de sujets de préoccupation, mais l’IA en ajoute de nouveaux. Étant donné que les systèmes propriétaires d’IA disponibles dans le commerce comportent le caractère biaisé des données et de l’algorithme sur lequel ils sont développés et s’accompagnent de restrictions et de problèmes de transparence, la fiabilité de l’IA propriétaire soulèvera inévitablement des questions quant à l’impartialité de ces systèmes.

Pourquoi l’impartialité est-elle importante pour les Nations unies ? Le principe d’impartialité est la clé de voûte de la crédibilité des Nations unies, garantissant que les conseils politiques restent objectifs, fondés sur des preuves et attentifs à la diversité des points de vue. Cette impartialité sera mise à l’épreuve lorsque les Nations unies réagiront à l’inévitable nécessité d’automatiser l’établissement de rapports, la rédaction et d’autres activités essentielles à leur fonctionnement.

Pour garantir l’impartialité, il faudrait que le cycle complet de l’IA soit transparent et explicable, depuis les données sur lesquelles reposent les modèles fondamentaux jusqu’à l’attribution d’un poids aux différents segments des systèmes d’IA.

Une approche inclusive du développement de l’IA est essentielle au respect du principe d’impartialité. Les trois premiers mots de la Charte des Nations unies, « Nous, les peuples », devraient guider le développement de l’IA à l’ONU. Les contributions des pays, des entreprises et des communautés du monde entier à AI@UN pourraient renforcer le potentiel élevé de l’IA pour soutenir les missions de l’ONU en matière de maintien de la paix mondiale, de promotion du développement et de protection des droits de l’Homme.

IA@ONU a deux objectifs principaux :

  • soutenir les discussions politiques sur la transformation durable de l’écosystème de l’ONU en matière d’IA ;
  • encourager les contributions des modèles et agents d’IA par les États membres et d’autres acteurs.

Comme point de départ, les principes directeurs suivants sont proposés pour le développement et le déploiement de modèles, de modules et d’agents d’IA au sein des Nations unies.

1. Logiciel libre : respecter les principes, les traditions et les pratiques de la communauté des logiciels libres. L’ouverture et la transparence devraient s’appliquer à toutes les phases et à tous les aspects du cycle de vie de l’IA, y compris la conservation des données et des connaissances pour les systèmes d’IA, la sélection des paramètres, et l’attribution de poids pour développer des modèles fondamentaux, des bases de données vectorielles, des graphes de connaissances et d’autres segments des systèmes d’IA. 

2. Modularité : développer des modules autonomes selon des normes et des paramètres communs. IA@ONU devrait commencer par des agents et des modules d’IA pour les activités et les opérations de base de l’ONU.

3. Bien public : faire la promotion du bien public en utilisant l’IA pour codifier les connaissances des Nations unies en tant que telles à l’usage des pays, des communautés et des citoyens du monde entier. Ce faisant, les Nations unies inspireraient la codification par l’IA de diverses sources de connaissances, y compris les textes anciens et la culture orale, en tant que patrimoine commun de l’humanité.   

4. Inclusion : permettre aux États membres, aux entreprises et aux universités de contribuer, par leurs capacités et leurs ressources, aux aspects techniques, aux connaissances et à la facilité d’utilisation d’IA@ONU.

5. Multilinguisme : représenter un large éventail de traditions linguistiques et culturelles. Une attention particulière devrait être accordée à la récolte des connaissances et de la sagesse disponibles dans les traditions orales qui ne sont pas disponibles dans le corpus écrit des livres et des publications.

6. Diversité : garantir la contribution d’un large éventail de perspectives professionnelles, générationnelles, culturelles et religieuses. Si IA@ONU doit viser à identifier les convergences entre les différents points de vue et approches, la diversité ne doit pas être supprimée par la démarche du plus petit dénominateur commun que l’on retrouve dans l’IA. La diversité doit être intégrée par la traçabilité transparente des sources à l’origine des résultats générés par l’IA. 

7. Accessibilité : adhérer aux normes les plus élevées en matière d’accessibilité, en particulier pour les personnes handicapées. IA@ONU doit accroître la participation des personnes handicapées aux activités de l’ONU, qu’il s’agisse de réunions ou de projets pratiques. Des solutions simples et une faible demande de bande passante devraient rendre le système abordable pour tous. 

8. Interopérabilité : aborder le problème des silos organisationnels dans la gestion des connaissances et des données au sein du système des Nations unies. L’interopérabilité devrait être facilitée par des ontologies et des taxonomies de la connaissance, ainsi que par la conservation des données et des normes techniques communes.

9. Professionnalisme : respecter les normes industrielles et éthiques les plus strictes en matière de planification, de codage et de déploiement d’applications logicielles. Pour ce faire, les solutions d’IA seront testées, évaluées et soumises à un processus d’examen par les pairs. L’objectif principal sera de maximiser le développement fiable de solutions d’IA afin d’avoir un impact direct sur la vie et le bien-être des êtres humains. 

10. Explicabilité : tracer chaque produit généré par l’IA, tel qu’un rapport ou une analyse, jusqu’aux sources utilisées par l’inférence de l’IA, y compris les textes, les images et les enregistrements sonores. L’explicabilité et la traçabilité garantiraient la transparence et l’impartialité des systèmes IA@ONU.

11. Protection des données et des connaissances : atteindre le plus haut niveau de protection des données, des connaissances et des autres entrées dans les systèmes d’IA. 

12. Sécurité : garantir le plus haut niveau possible de sécurité et de fiabilité dIA@ONU. Les logiciels libres, l’équipe rouge (red-teaming) et d’autres méthodes garantiront la protection des systèmes en permettant au plus grand nombre possible d’yeux critiques de tester et d’évaluer le code et les algorithmes de l’IA. Les communautés de l’IA seront encouragées à contribuer au red-teaming (l’équipe rouge) et à d’autres tests du système IA@ONU.

13. Durabilité : la réalisation des ODD et de l’Agenda 2030 par le biais de trois approches principales : premièrement, veiller à ce que les ODD bénéficient d’un poids plus important dans le développement de modèles et d’outils d’IA ; deuxièmement, les systèmes d’IA eux-mêmes devraient être durables, par exemple en partageant le code, en développant des ressources, et en fournissant des documentations et des pistes de développement appropriées ; troisièmement, les solutions d’IA devraient être développées et déployées en gardant à l’esprit la durabilité de l’environnement.

14. Capacité : en développant un système d’IA, l’ONU devrait accroître ses propres capacités en la matière et celles d’un plus grand nombre de personnes. Le développement des capacités devrait être : (a) holistique, impliquant le Secrétariat de l’ONU, les représentants des États membres et d’autres communautés impliquées dans les activités de l’ONU ; et (b) complet, couvrant un large éventail de capacités d’IA, allant d’une compréhension de base de l’IA à des compétences techniques de haut niveau. 

15.  Protection de l’avenir : planifier et déployer des systèmes adaptés aux futures tendances technologiques. L’expérience et l’expertise acquises dans le cadre d’IAI@ONU devraient être utilisées pour traiter d’autres technologies émergentes, telles que la réalité augmentée/virtuelle et l’informatique quantique. 

Opportunités en cas de crises : la transformation de l’IA provoquera inévitablement des tensions en raison de son impact sur les couches les plus profondes du fonctionnement des Nations unies. L’opposition probable fondée sur la peur humaine et l’attachement au statu quo devrait être ouvertement abordée et recadrée autour des opportunités que la transformation de l’IA ouvrira aux niveaux individuel et institutionnel.

Par exemple, l’IA peut aider les petits pays et les pays en développement à participer de manière plus informée et plus efficace aux travaux des Nations unies. L’IA peut contribuer à compenser la taille réduite de leurs missions et services diplomatiques, qui doivent suivre la même dynamique diplomatique que les systèmes plus importants. L’accent mis sur l’IA réduit l’asymétrie actuelle de l’IA.

L’IA peut également aider le Secrétariat des Nations unies à recentrer son temps et ses ressources, et à consacrer moins de temps à la bureaucratie traditionnelle, comme la préparation des rapports, afin de pouvoir travailler davantage sur le terrain dans les États membres, où leur aide est cruciale.

Prochaines étapes : s’engager sur la voie de l’intégration de l’IA dans les opérations des Nations unies n’est pas simplement une étape, mais un saut dans l’avenir, qui exige de l’audace, un esprit de coopération et un dévouement inébranlable aux idéaux qui ont ancré les Nations unies depuis leur création. Le potentiel de l’IA pour soutenir la mission de l’ONU de maintenir la paix dans le monde, de faire progresser le développement et de défendre les droits de l’Homme est immense. En fait, la nécessité d’adopter un cadre d’IA à source ouverte dépasse le besoin d’innovation technologique. En adoptant une approche ouverte de l’IA, les Nations unies seront en mesure d’évoluer, de prendre l’initiative et de rester pertinentes dans un paysage mondial en mutation rapide.

En tirant parti du pouvoir de transformation de l’IA, les Nations unies peuvent transformer un défi imminent en un moment décisif, garantissant ainsi la pertinence et le leadership de l’organisation dans la définition de la voie du progrès humain pour tous.

Ce texte a été adapté de AI@UN : Navigating the tightrope between innovation and impartiality, publié pour la première fois sur le blogroll de Diplo.

Analyse

Comment les dialogueurs d’IA maîtrisent le langage : les enseignements de la linguistique de Saussure

La linguistique, entremêlée à la technologie moderne, suscite des questions sur la façon dont les dialogueurs fonctionnent et répondent de manière cohérente à diverses entrées. Les dialogueurs, alimentés par de grands modèles de langage (LLM), comme ChatGPT, acquièrent une cognition numérique et formulent des réponses en utilisant des principes enracinés dans les théories linguistiques de Ferdinand de Saussure.Les travaux de Saussure, au début du XXe siècle, ont jeté les bases de la compréhension du langage par la syntaxe et la sémantique. La syntaxe désigne les règles qui régissent l’agencement des mots pour former des phrases significatives. Saussure considérait la syntaxe comme un système de conventions au sein d’une communauté linguistique, lié à d’autres éléments linguistiques tels que la sémantique. La sémantique implique l’étude du sens dans la langue. Saussure a introduit le concept de signe, composé du signifiant (son/image) et du signifié (concept), qui est crucial pour comprendre comment les LLM traitent et interprètent le sens des mots.

 Art, Modern Art, Graphics, Painting, Person, Tape, Face, Head

Comment les LLM traitent le langage. Les LLM tels que ChatGPT traitent et comprennent le langage par le biais de plusieurs mécanismes fondamentaux.

  1. Entraînement sur de grandes quantités de données textuelles provenant d’Internet pour prédire le mot suivant dans une séquence.
  2. Segmentation en unités pour diviser le texte en éléments plus petits.
  3. Apprentissage des relations entre les mots et les phrases pour la compréhension sémantique.
  4. Utilisation de représentations vectorielles pour reconnaître les similitudes et générer des réponses adaptées au contexte.
  5. Exploitation de l’architecture des transformateurs pour traiter efficacement les contextes longs et les structures linguistiques complexes.

Les LLM transforment le texte en unités segmentées (signifiants) et les mettent en correspondance avec des enchâssements qui capturent leurs significations (signifiés). Le modèle apprend ces enchâssements en traitant de grandes quantités de texte, en identifiant des modèles et des relations analogues aux structures linguistiques de Saussure.

Sémantique et syntaxe dans les grands modèles de langage (LLM). La compréhension et la production de texte dans les LLM impliquent un traitement à la fois sémantique et syntaxique. 

Pour traiter la sémantique, les LLM s’appuient sur (a) des enchâssements de mots contextuels qui saisissent le sens des mots dans différents contextes en fonction de leur utilisation, (b) un mécanisme d’attention qui donne la priorité aux mots importants, et (c) une compréhension contextuelle en couches qui traite les mots ayant plusieurs sens apparentés (polysémie) et différents mots ayant le même sens (synonymie). Le modèle est pré-entraîné sur des modèles de langage généraux et affiné sur des ensembles de données spécifiques pour une meilleure compréhension sémantique. 

Concernant la syntaxe, les LLM utilisent (a) l’encodage positionnel pour comprendre l’ordre des mots, (b) les mécanismes d’attention pour maintenir la structure de la phrase, (c) le traitement en couches pour construire des phrases complexes, et (d) l’apprentissage de la grammaire probabiliste à partir de grandes quantités de texte. La segmentation et la modélisation des séquences permettent de suivre les relations entre les mots, et le modèle de transformation intègre à la fois la structure et le sens de la phrase à chaque étape, garantissant ainsi que les réponses sont à la fois significatives et grammaticalement correctes. L’entraînement sur divers ensembles de données améliore encore sa capacité à généraliser les différentes façons d’utiliser le langage, faisant du Dialogueur un puissant outil de traitement du langage naturel.

L’intégration des théories linguistiques de Saussure aux mécanismes cognitifs des grands modèles de langage éclaire le fonctionnement interne des systèmes d’IA contemporains et renforce également la pertinence des théories linguistiques classiques à l’ère de l’IA.

Ce texte a été adapté de Au commencement était le mot, et le mot était avec le Dialogueur, et le mot était le Dialogueur, publié pour la première fois sur le site web du Digital Watch Observatory.

En Bref

Les géants des médias sociaux obtiennent gain de cause dans l’affrontement sur la liberté d’expression devant la Cour suprême des États-Unis

Les plateformes de médias sociaux jouent un rôle prépondérant dans la vie des gens, non seulement pour la communication, mais aussi pour la réception et la diffusion d’informations. Dans le même temps, le contenu des médias sociaux peut présenter certains risques, tels que la possibilité de discours haineux, la diffusion de fausses informations et de désinformation, et le harcèlement.

Cette situation a soulevé des questions sur leurs responsabilités dans la régulation de ces contenus, ainsi que sur le rôle des gouvernements dans la prise de mesures.

Les plateformes de médias sociaux ont-elles le droit de s’exprimer librement ? Les gouvernements peuvent-ils mettre en œuvre des politiques à leur encontre et à l’encontre de leurs propres politiques de contenu ? La Cour suprême des États-Unis s’est penchée sur ces questions dans ses arrêts Moody vs NetChoice et NetChoice, LLC vs Paxton.

NetChoice et la Computer and Communications Industry Association (CCIA), une association réunissant des entreprises de médias sociaux et des plateformes Internet, ont contesté les lois de deux États américains, la Floride et le Texas. Ces lois ont été adoptées en 2021 alors que le Parti républicain critiquait de plus en plus l’application par les entreprises de médias sociaux de leurs propres politiques. Les tensions se sont aggravées lorsque de grandes plateformes comme Twitter, YouTube et Meta, de Facebook, ont suspendu les comptes de l’ancien président Trump à la suite de ses remarques sur l’émeute du 6 janvier 2021 au Capitole.

NetChoice et CCIA ont affirmé que les lois de la Floride et du Texas violaient les droits du premier amendement des entreprises privées et que les gouvernements ne devraient pas être autorisés à intervenir dans les politiques d’expression des entreprises privées. Un groupe de politologues a déposé un mémoire d’amicus curiae indiquant que ces deux lois ne fixent pas de seuil quant à ce qu’elles considèrent comme un discours haineux, et quant aux discours dangereux et violents liés aux élections qui pourraient empêcher les plateformes de médias sociaux de modérer les menaces à l’encontre des responsables électoraux.

Par ailleurs, les représentants du Texas et de la Floride affirment que ces lois visent à réglementer la responsabilité des plateformes de médias sociaux plutôt qu’à restreindre le discours en ligne, tout en soulignant que le premier amendement ne s’applique pas aux entreprises privées. Une cour d’appel fédérale américaine a invalidé la loi de la Floride, tandis qu’une autre a confirmé la loi du Texas. Toutefois, les deux lois ont été suspendues dans l’attente de la décision finale de la Cour suprême des États-Unis.

 Indoors, Courtroom, Room

La Cour suprême a décidé que les décisions des tribunaux de première instance n’étaient pas suffisantes pour garantir le droit à la liberté d’expression en vertu du premier amendement et que les deux lois étaient inconstitutionnelles.

En conclusion de sa décision, la Cour suprême a estimé que les plateformes de médias sociaux sont protégées par le premier amendement lorsqu’elles créent du contenu. La Cour suprême a également statué que la présentation d’une collection de discours d’autrui est considérée comme une activité expressive. La Cour suprême a notamment déclaré que :

« Dans la mesure où les plateformes de médias sociaux créent des produits expressifs, elles bénéficient de la protection du premier amendement. Et bien que ces affaires soient ici à un stade préliminaire, le dossier actuel suggère que certaines plateformes, au moins dans certaines fonctions, sont engagées dans une activité d’expression. En outre, “l’activité expressive inclut la présentation d’une compilation de discours créés à l’origine par d’autres”. »

Pour l’essentiel, cette décision crée un précédent en établissant des droits à la liberté d’expression au titre du premier amendement pour les plateformes de médias sociaux et les entreprises privées aux États-Unis. En effet, les États américains ne peuvent pas mettre en œuvre des politiques limitant leur capacité à réglementer le contenu diffusé sur leurs plateformes. Cela pourrait empêcher les gouvernements de promulguer des lois conduisant à la perte d’indépendance des plateformes de médias sociaux dans la régulation de leur contenu.

Les gouvernements avancent à grands pas dans la surveillance du marché de la concurrence numérique

En 1996, John Perry Barlow a rédigé une « Déclaration d’indépendance du cyberespace ». Ce document d’anthologie, qui reflétait la culture libertaire de l’internet de l’époque, s’opposait à l’intervention gouvernementale et à la réglementation du secteur technologique en plein essor. En conséquence, les gouvernements du monde entier ont adopté une approche non interventionniste, partant du principe que la réglementation risquait d’étouffer l’innovation.

Près de 30 ans plus tard, cette vision des choses a radicalement changé. Ces dernières années, des rapports publiés par plusieurs organisations, telles que la Banque mondiale, l’Internet Society et la CNUCED, ont montré une concentration croissante de la richesse et du pouvoir dans l’économie numérique. Les fractures liées aux données sont particulièrement pertinentes dans ce contexte, car elles entraînent une concentration en amont, dans les secteurs technologiques à forte intensité de données, tels que l’IA. Dans ce contexte, les enquêtes sur les comportements potentiellement anticoncurrentiels adoptés par les entreprises technologiques se multiplient.

Dans l’UE, des enquêtes récentes ont abouti à la première accusation portée par la Commission européenne contre une entreprise technologique en vertu de la loi sur les marchés numériques (Digital Markets Act, DMA), une loi conçue pour limiter la domination des grandes entreprises technologiques et favoriser une concurrence loyale. Selon les conclusions préliminaires d’une enquête lancée en mars, Apple serait en infraction avec la loi sur les marchés numériques. L’App Store, d’Apple, évincerait les places de marché concurrentes en rendant plus difficile pour les utilisateurs le téléchargement d’applications à partir d’autres magasins, et en ne permettant pas aux développeurs d’applications de communiquer librement et de conclure des contrats avec leurs utilisateurs finaux. Apple a eu la possibilité d’examiner les conclusions préliminaires et peut encore éviter une amende si elle présente une proposition satisfaisante pour résoudre le problème.

Certains pays durcissent également leur législation en matière de concurrence. L’« effet Bruxelles » et l’influence de la DMA sont visibles dans le projet de loi sur la concurrence numérique, proposé par le gouvernement indien pour compléter les lois antitrust existantes. Comme la DMA, la loi viserait les grandes entreprises et pourrait prévoir des amendes aussi lourdes. Il serait notamment interdit aux géants de la technologie d’exploiter les données non publiques des utilisateurs et de favoriser leurs propres produits ou services sur leurs plateformes. Il leur serait également interdit de restreindre la capacité des utilisateurs à télécharger, installer ou utiliser des applications tierces.

Le projet de loi suscite l’inquiétude des entreprises technologiques. Un groupe de défense américain s’y est opposé, craignant son incidence sur les entreprises. Inspirées par la croyance commune qui a dominé le secteur technologique dans les années 1990, les entreprises technologiques affirment que le projet de loi indien pourrait étouffer l’innovation. Toutefois, il est peu probable que cette affirmation prospère. 

Les inquiétudes concernant la concurrence dans le secteur technologique se font de plus en plus vives aux États-Unis, traditionnellement partisans d’une réglementation minimale. Les États-Unis renforcent les contrôles de l’industrie de l’IA, le DOJ et la Federal Trade Commission (FTC) se partageant la surveillance : la FTC réglementera OpenAI et Microsoft, tandis que le DOJ supervisera Nvidia. Bien que moins actifs que l’UE en matière de réglementation de la concurrence, les États-Unis surveillent de près les fusions et les acquisitions. Ce récent accord entre les deux organes gouvernementaux a ouvert la voie au lancement d’enquêtes sur la concurrence.

La concurrence devient de plus en plus un terrain de jeu où l’activité et la surveillance des gouvernements sont importantes. Alors que les pays réaffirment leur compétence, les revendications de cyberindépendance semblent être un lointain écho du passé.

Ce texte a été publié pour la première fois sur le blog de Diplo. Lire la version originale.


Actualités de la Francophonie

 Logo, Text

Francophones, hispanophones et lusophones plaident pour plus de diversité linguistique et culturelle dans le PNM

A l’initiative de l’OIF et du Groupe des Ambassadeurs francophones de New York, un atelier de discussion sur la diversité culturelle et linguistique à l’ère du numérique et des technologies émergentes s’est tenu aux Nations unies le 18 juin 2024, dans le contexte des négociations du Pacte numérique mondial. 

L’évènement organisé en partenariat avec les Représentations permanentes à l’ONU de la République démocratique du Congo – qui préside le GAF-NY, du Brésil – qui coordonne la Communauté des pays de langue portugaise, du Mexique – membre du Groupe des amis de l’Espagnol, de la Roumanie et du Cabo Verde, a permis d’alerter sur les menaces et de mettre en lumière le potentiel des développements numériques pour la préservation, la promotion et le respect de la diversité des langues et des cultures.  

Les Ambassadeurs présents ont collectivement réaffirmé leur engagement en faveur de l’inclusion numérique, mettant en garde contre le risque d’élargir les fractures numériques à travers un Pacte qui n’adresserait qu’insuffisamment les défis liés à la diversité. De nombreuses idées ont été partagées sur les moyens de renforcer les formulations relatives à la diversité culturelle et linguistique dans le texte du Pacte.

L’OIF avait convié plusieurs experts à apporter leur point de vue. Ainsi, le Président de l’Observatoire de la diversité linguistique et culturelle dans l’Internet, Daniel Pimienta, est revenu sur la présence et les évolutions des langues sur la toile. Pour sa part, Hannah Taieb, Directrice du développement chez Spideo, entreprise spécialisée dans la fourniture d’algorithmes de recommandation sur internet, a replacé l’enjeu de la découvrabilité des contenus culturels dans un contexte de prolifération des plateformes numériques, dont l’offre de plus en plus diversifiée est guidée par l’expérience utilisateur. Elle a ainsi plaidé pour plus de transparence des plateformes sur le fonctionnement de leurs algorithmes de recommandation.

 People, Person, Accessories, Formal Wear, Tie, Indoors, Adult, Male, Man, Computer, Electronics, Laptop, Pc, Chair, Furniture, Crowd, Glasses, Architecture, Building, Classroom, Room, School, Audience, Lecture, Computer Hardware, Hardware, Monitor, Screen

Cet atelier de haut niveau s’inscrit dans l’action multiforme que l’OIF mène pour appuyer ses pays membres engagés dans les négociations actuelles du Pacte numérique mondial à New York. Ainsi, la Représentation 

de l’OIF à Genève a fait circuler des analyses détaillées des propositions de texte après qu’aient été publiées la 1e révision puis la 2e révision du document constituant le Pacte numérique. Ces analyses, complétées par le compte-rendu des prises de parole des délégations lors des sessions se déroulant à New York, permettent aux pays francophones d’avoir une bonne compréhension du contexte et des enjeux derrière les évolutions des formulations.

Les experts numériques francophones de Genève échangent avec M. Zavazava, Directeur du bureau du développement des télécommunications de l’Union internationale des Télécommunications (UIT)

La Représentation permanente de l’OIF à Genève (RPGV) a hébergé le 21 juin dans ses locaux une rencontre entre les experts numériques des missions diplomatiques francophones à Genève et M. Cosmos Zavazava, Directeur du Bureau du développement des télécommunications à l’Union internationale des Télécommunications (UIT) 

Ce dernier a présenté les actions de son bureau en faveur des pays en développement et notamment des pays francophones. La question de l’aide que l’UIT peut apporter en matière de transition numérique et de développement des infrastructures d’intelligence artificielle pour combler le fossé numérique existant entre pays, a été particulièrement évoquée.

 People, Person, Indoors, Adult, Male, Man, Crowd, Chair, Furniture, Accessories, Bag, Handbag, Lamp, Female, Woman, Electronics, Speaker, Computer, Laptop, Pc, Formal Wear, Tie, Clothing, Footwear, Shoe, Audience, Computer Hardware, Hardware, Monitor, Screen, Hall, Face, Head, Glasses

Déclaration conjointe du Groupe des Ambassadeurs Francophones (GAF) de Genève autour des technologies émergentes lors de la session du Conseil des droits de l’Homme

Les coordinations thématiques « Droits de l’Homme » et « numérique » du GAF de Genève ont travaillé de concert pour élaborer une Déclaration conjointe qui a été lue par l’Ambassadeur de Belgique, S.E.M. Marc Pecsteen de Buytswerve, au nom du GAF dans le cadre du Dialogue interactif sur le rapport du Haut-Commissaire aux droits de l’Homme « Les technologies nouvelles et émergentes »  qui s’est tenu lundi 1er juillet à Genève.

Cette Déclaration a mis en relief le soutien des pays francophones au développement d’un cadre multilatéral de gouvernance des technologies numériques et de l’intelligence artificielle, sur la base d’une approche multidimensionnelle et multi-acteurs, et basée sur les droits humains, l’universalité ainsi que la diversité culturelle et linguistique.

 People, Person, Adult, Female, Woman, Crowd, Architecture, Building, College, Male, Man, Chair, Furniture, Indoors, Electrical Device, Microphone, Computer Hardware, Electronics, Hardware, Monitor, Screen, Audience, Speaker, Accessories, Formal Wear, Tie, Classroom, Room, School, Face, Head, Lecture

Ateliers en collaboration avec l’UNIDIR autour des enjeux d’IA dans le domaine militaire et des cybermenaces

La Représentation de l’OIF à Genève a lancé avec l’Institut des Nations unies pour la recherche sur le désarmement (UNIDIR) et la Mission permanente de la France auprès de la Conférence pour le désarmement, une série de six séances d’information en langue française sur des questions liées au désarmement. La première de ces séances, organisée le 19 juin, a porté sur « l’intelligence artificielle dans le domaine militaire : enjeux et gouvernance », Ont été examinées les capacités et potentialités de l’IA appliquées au domaine militaire, les bénéfices et les risques qu’elles entraînent, les questions juridiques et éthiques qu’elles soulèvent, ainsi que la gouvernance qui pourrait être établie pour assurer le développement d’une IA maitrisée et responsable. Puis le 24 juin, c’est la thématique des cybermenaces, son paysage et les évolutions récentes, notamment dans le contexte de conflits armés, qui a fait l’objet d’échanges animés lors du deuxième atelier de cette série.

 People, Person, Adult, Female, Woman, Audience, Crowd, Indoors, Lecture, Room, Seminar, Male, Man, Architecture, Building, Classroom, School, Electronics, Screen, Computer Hardware, Hardware, Monitor, Electrical Device, Microphone, Mobile Phone, Phone

DW Weekly #168 – 12 July 2024

 Page, Text

Dear readers, 

Welcome to another issue of the Digital Watch weekly! 

We will spotlight another round of the China vs the USA tech competition. Last week, WIPO published a report on the landscape of patents in generative AI or GenAI (think AI that can generate images, videos, text, music, code, etc.) One of the main takeaways from this report, which has been all over the headlines, is that China has a substantial lead in AI patents, with six times more GenAI inventions than the second country on the list, the USA. The Republic of Korea, Japan, and India round up the top five.

Titled 'Number of GenAI patent applications, a stepped award stage shows flags representing the 5 top contenders: (1) China, (2) the USA, (3) South Korea, (4) Japan, and (5) India

This week, a survey of business decision-makers has shown that 83% of Chinese organisations are using GenAI

Does all this recent data show that China is winning the AI race? Not quite. While the country is ahead in some areas, the USA is a strong contender. The USA produced more notable machine-learning models in 2023 (61 compared to China’s 15), leads in AI foundation models, variational autoencoders (VAEs), and private AI investments. Plus, we should remember that the number of patents doesn’t correlate to their quality or impact – the mere registration of patents doesn’t automatically translate into apps or services being put on the market. 

(Sidebar: foundation models are machine learning models trained on broad and diverse data and which can be adapted to a wide range of tasks across different domains; they serve as a base for more specialised models dedicated to specific tasks or areas. A variational autoencoder (VAE) is a generative AI algorithm that uses deep learning to generate new content based on the structure of the input data.) 

The USA nudges ahead with 24% of organisations fully implementing GenAI compared to 19% in China. In addition, since 2018, the USA has been embroiled in a chip war with China, blacklisting Chinese chipmakers and tightening controls on the export of its most advanced chip-related technologies. This is particularly important because there are almost no semiconductors without some kind of US-trademarked bits in their design or production processes, making things tricky for Chinese chipmakers and industries reliant on such technology, including AI, in the meantime.

Stepping away from the US-China competition, it’s important to acknowledge that there are other actors out there to watch. One is India, which had the highest growth rates in GenAI patent family publications (56% per year). 

Interestingly, another survey shows that most (polled) Americans view winning the AI race against China as secondary to a cautious approach to AI development to prevent its misuse by adversaries. Only 23% believe the USA should rapidly build powerful AI to outpace China and gain a decisive advantage.

While the USA and China are grappling with each other, and others are rushing to catch up, the contender closest to the crown right now is AI itself.

The discussions at the 8th substantive session of the Open-ended Working Group (OEWG) on security of and in the use of information and communications technologies 2021–2025 are still ongoing, today being the last day. AI reports and transcripts are available on our dedicated web page, with a human-generated analysis planned for next week.

Andrijana and the Digital Watch team


Highlights from the week of 5-12 July 2024

building 1011876 1280

The decision comes amid regulatory scrutiny from antitrust watchdogs in Europe, the UK, and the US concerning Microsoft’s potential influence over OpenAI.

the white house

President Joe Biden has formed a team of experts to create standards for AI training and deployment across industries.

serbia ai supercomputer

The new AI Development Strategy 2024-2030 builds on this solid foundation, aiming to nurture a vibrant AI ecosystem in Serbia.

eu cybersecurity standards

The European Commission’s AI Act will classify AI-based cybersecurity and emergency services in connected devices as high-risk, requiring extensive testing and security measures, significantly impacting various sectors like medical devices…

OpenAi

OpenAI announced it would block Chinese users from accessing its services on 9 July, amid rising US-China tensions, affecting developers who relied on OpenAI tools.

electronics 4972649 1280

The semiconductor industry, now leading the S&P 500, is experiencing explosive demand driven by AI advancements.

japanese and us flags semiconductor consortium

Based in Silicon Valley, the consortium will focus on developing advanced back-end technologies for semiconductor packaging, aiming to be fully operational next year.

eu and china flags

The initiative comes amid rising EU-China tensions, exemplified by tariffs of up to 37.6% on Chinese electric vehicles.

woman using laptop wood desk with cyberattack warning screen cyber security concept

US authorities have disrupted a sophisticated Russian disinformation campaign, Meliorator, which uses AI to create fake social media personas and spread false information in the US and internationally.

hacker working in the darkness

China’s embassy in Australia dismissed the allegations as ‘political manoeuvring’.

nato flag north atlantic treaty organization flag waving

The move, unveiled during the 2024 NATO Summit in Washington, DC, marks NATO’s 75th anniversary and addresses the increasing cyber threats, especially following the Russian invasion of Ukraine in 2022.

print at home concert tickets

Hackers leaked nearly 39,000 print-at-home tickets for major events in an extortion scheme against Ticketmaster.

new zealand

The shift aims to ensure long-term success and maintain a multistakeholder approach involving governments, tech companies, and civil society.

taiwan flag is depicted on the screen with the program code

Despite this slowdown, the bank plans to conduct public hearings next year to inform the public about digital currency.

amazon logo

Amazon is under pressure from the European Commission to enhance transparency and ensure compliance with new regulations to tackle illegal and harmful content online.

flag of sri lanka

The legislative change introduces three new types of licences for satellite internet service providers, one of which allows Starlink to apply as a licensed service provider pending regulatory approval.



ICYMI

The Detroit Police Department has agreed to new rules limiting how it can use facial recognition technology, prompted by a lawsuit over a wrongful arrest in 2020. The EU and France flex their antitrust powers against Microsoft, Apple and Nvidia. Watch for details!


Upcoming

8N6ikeWA un flag blue

The UN OEWG 2021–2025 8th substantive session will focus on adopting the group’s annual progress report (APR), taking stock of the group’s discussions over the previous year and charting the way forward for the group. The GIP follows the event; just-in-time reports and transcripts are available on our dedicated web page.

440 360 max July newsletter banner

In the July issue of our monthly newsletter, we look at how the UN can navigate the tightrope between innovation and impartiality when integrating AI in its operations, explore how AI chatbots master language using principles rooted in the linguistic theories of Ferdinand de Saussure, look at recent governments’ actions on digital antitrust oversight, and explain how social media giants won in a free speech showdown at the US Supreme Court.

Digital Watch newsletter – Issue 91 – July 2024

Front page of the Newsletter

Snapshot: The developments that made waves

AI governance

The UN General Assembly has adopted a non-binding resolution on AI capacity building, led by China, to enhance developing countries’ AI capabilities through international cooperation. It also calls for support from international organisations and financial institutions. African ICT and communications ministers have endorsed the Continental AI Strategy and the African Digital Compact to boost the continent’s digital transformation. The G7 Leaders’ Communiqué emphasised a coordinated strategy for handling AI’s opportunities and challenges, introducing an action plan for workplace AI adoption and underlining initiatives such as advancing the Hiroshima Process International Code of Conduct, supporting SMEs, and promoting digital inclusion and lifelong learning.

The International Monetary Fund has recommended fiscal policies for governments grappling with the economic impacts of AI, including taxes on excess profits and a carbon levy.

China leads the world in generative AI patent requests, significantly outpacing the USA. At the same time, US tech companies dominate in producing cutting-edge AI systems, according to the World Intellectual Property Organization (WIPO). A European Commission report shows the EU lags behind its 2030 AI targets, with only 11% of enterprises using designated AI technologies, far short of the 75% target. The Japanese Defence Ministry has introduced its first AI policy to enhance defence operations. Brazil is partnering with OpenAI to modernise legal processes, reduce court costs, and improve efficiency in the solicitor general’s office.

Technologies

The USA has introduced draft rules to regulate investments in China, focusing on AI and advanced technology sectors that may pose national security threats. The USA plans to expand sanctions on semiconductor chips and other goods sold to Russia, targeting Chinese third-party sellers. Discussions are ongoing with the Netherlands and Japan to restrict 11 Chinese chipmaking factories and extend equipment export controls. The USA faces a projected shortage of 90,000 semiconductor technicians by 2030, prompting the Biden administration to launch a workforce development program.

The European Commission is seeing industry views about China’s increased production of older-generation computer chips.

China will develop standards for brain-computer interfaces (BCI) through a new technical committee, focusing on data encoding, communication, visualisation, electroencephalogram data collection, and applications in various fields.

Infrastructure

Telecommunications companies from Kazakhstan and Azerbaijan will invest over USD 50 million in laying 370 kilometres of fibre optic cables under the Caspian Sea. Meanwhile, Senegal’s new digital chief announced plans to enhance digital infrastructure, coordinate government programs, foster collaborations, and build on previous achievements to increase the digital economy’s GDP contribution.

Cybersecurity

The UN Security Council held an open debate on cybersecurity, focusing on evolving cyber threats and the need for positive digital advancements.

A recent cyberattack on the cloud storage company Snowflake is shaping up to be one of the largest data breaches ever, impacting hundreds of Snowflake business customers and millions of individual users. Indonesia’s national data centre was hit by a variant of LockBit 3.0 ransomware, disrupting immigration checks and public services. The hackers have since apologised and offered to release the keys to the stolen data. The University Hospital Centre in Zagreb, Croatia, also suffered a cyberattack by LockBit. Despite rising ransomware attacks, a Howden report indicates that global cyber insurance premiums are decreasing as businesses improve their loss mitigation capabilities. Additionally, nearly ten billion unique passwords were leaked in a collection named RockYou2024, heightening risks for users who reuse passwords.

Australia has mandated internet companies to create enforceable codes within six months to prevent children from accessing inappropriate content. New Zealand transitioned the Christchurch Call to Action against online terrorist content into an NGO, now funded by tech companies like Meta and Microsoft.

Digital rights

The EU’s proposed law mandating AI scans of messaging app content to detect child sexual abuse material (CSAM) faces criticism over privacy threats and potential false positives. EU regulators charged Meta with breaching tech rules via a ‘pay or consent’ ad model on Facebook and Instagram, alleging it forced users to consent to data tracking. The US Department of Justice (DOJ) plans a lawsuit against TikTok for alleged children’s privacy violations. Google is accused by European data protection advocacy group NOYB (none of your business) of tracking users without their informed consent through its Privacy Sandbox feature.

Legal

The International Criminal Court is investigating alleged Russian cyberattacks on Ukrainian infrastructure as potential war crimes. In Australia, legal action has been initiated against Medibank for a data breach affecting 9.7 million individuals. ByteDance and TikTok are challenging a US law aiming to ban the app, citing concerns about free speech. Global streaming companies are contesting new Canadian regulations mandating 5% of revenues be used for local news, questioning the legality of the government’s actions.

Internet economy

China’s Ministry of Commerce has introduced draft rules to bolster cross-border e-commerce by promoting the establishment of overseas warehouses and improving data management and export supervision. Nvidia is facing potential charges in France over allegations of anti-competitive behaviour. The first half of 2024 saw a significant surge in cryptocurrency theft, with over USD 1.38 billion stolen by 24 June.

Development

The first part of the Broadband Commission’s annual State of Broadband report ‘Leveraging AI for Universal Connectivity’ explores AI’s impact on e-government, education, healthcare, finance, and environmental management, and its potential to bridge or widen the digital divide. The second part will provide further insights into AI’s development and propose strategies for equitable digital advancement. India will require USB-C as the standard charging port for smartphones and tablets starting in June 2025, aligning with the EU’s efforts to enhance user convenience and reduce electronic waste.

Sociocultural

New York state lawmakers passed a law limiting social media platforms in displaying addictive algorithmic content to users under 18 without parental consent. The European Commission has asked Amazon for details on how it complies with Digital Services Act rules, focusing on transparency in its recommender systems. Google Translate is significantly expanding, adding 110 languages, driven by AI advancements.

THE TALK OF THE TOWN – GENEVA

From 4 to 14 June, the Council of the International Telecommunication Union (ITU)  made key decisions on space development, green digital action, and global digital cooperation. The council reviewed the ITU Secretary-General’s report on the implementation of the Space 2030 Agenda, focusing on leveraging space technology for sustainability. Resolutions were drafted to highlight ITU’s role in using digital technologies for sustainability, with a report on current green digital initiatives. ITU will continue engaging with the Global Digital Compact (GDC) to enhance global digital cooperation.

On 14 June, the first UN Virtual Worlds Day showcased technologies like virtual and augmented reality, the metaverse, and spatial computing to advance SDGs. The event included a high-level segment, real-world applications, discussions on policy, and the launch of the Global Initiative on Virtual Worlds – Discovering the CitiVerse, a platform to develop frameworks, raise awareness, share best practices, and test metaverse solutions in cities.


AI@UN: Navigating the tightrope between innovation and impartiality

The UN is not short on risks, but AI adds novel ones for the organisation. As off-the-shelf AI proprietary systems carry the bias of data and the algorithm on which it is developed and come with limitations and challenges for transparency, reliance on proprietary AI will open inevitable questions about the impartiality of such systems.

Why is impartiality important for the UN? The principle of impartiality is the linchpin of the UN’s credibility, ensuring that policy advice remains objective, grounded in evidence, and sensitive to diverse perspectives. This impartiality will be tested as the UN reacts to the inevitable need to automate reporting, drafting, and other core activities central to its operation. 

Ensuring impartiality would require transparency and explainability of the full AI cycle, from the data on which foundational models are based to assigning weights to different segments of AI systems.

An inclusive approach to AI development is key for upholding the principle of impartiality. We, the peoples, the first three words of the UN Charter,  should guide the development of AI at the UN. Contributions of countries, companies, and communities worldwide to AI@UN could bolster the high potential of AI to support the UN’s missions of upholding global peace, advancing development, and protecting human rights.

AI@UN has two main goals:

  • support policy discussions on the sustainable AI transformation of the UN ecosystem
  • inspire the contributions of AI models and agents by member states and other actors
Emblem of the UN in white on a blue disk superimposed on an AI circuit with connectors radiating outwards from the centred disc.

As a starting point, the following guiding principles are proposed for the development and deployment of AI models, modules, and agents at the UN: 

1. Open source: Abiding by the open-source community’s principles, traditions, and practices. Openness and transparency should apply to all phases and aspects of the AI life-cycle, including curating data and knowledge for AI systems, selecting parameters and assigning weights to develop foundational models, vector databases, knowledge graphs, and other segments of AI systems. 

2. Modularity: Developing self-contained modules according to shared standards and parameters. AI@UN should start with AI agents and modules for core UN activities and operations.

3. Public good: Walking the talk of public good by using AI to codify UN knowledge as a public good to be used by countries, communities, and citizens worldwide. By doing so, the UN would inspire the AI-enabled codification of various knowledge sources, including ancient texts and oral culture, as the common heritage of humankind.   

4. Inclusivity: Enabling member states, companies, and academia to contribute, by their capacities and resources, to the technical, knowledge, and usability aspects of AI@UN. 

5. Multilingualism: Representing a wide range of linguistic and cultural traditions. A special focus should be on harvesting the knowledge and wisdom available in oral traditions that are not available in the written corpus of books and publications.

6. Diversity: Ensuring inputs from a wide range of professional, generational, cultural, and religious perspectives. While AI@UN should aim to identify convergences between different views and approaches, diversity should not be suppressed by the least common denominator approach inferred in AI. Diversity should be built in through the transparent traceability of sources behind AI-generated outputs. 

7. Accessibility: Adhering to the highest standards for accessibility, in particular for people with disabilities. AI@UN must increase the participation of people with disabilities in UN activities, from meetings to practical projects. Simple solutions and low-bandwidth demand should make the system affordable for all. 

8. Interoperability: Addressing the problem of organisational silos in managing knowledge and data within the UN system. Interoperability should be facilitated by knowledge ontologies and taxonomies, data curation, and shared technical standards.

9. Professionalism: Following the highest industry and ethical standards of planning, coding, and deploying software applications. This will be achieved by testing, evaluating, and submitting AI solutions to a peer-review process. The main focus will be maximising the reliable development of AI solutions to directly impact human lives and well-being. 

10. Explainability: Tracing every AI-generated artefact, such as a report or analysis, to sources used by AI inference, including texts, images and sound recording. Explainability and traceability would ensure transparency and impartiality of AI@UN systems.

11. Protection of data and knowledge: Achieving the highest level in protecting data, knowledge and other inputs into AI systems. 

12. Security: Guaranting the highest possible level of security and reliability of AI@UN. Open source, red-teaming, and other approaches will ensure that the systems are protected by having as many critical eyes as possible to test and evaluate AI code and algorithms. AI communities will be encouraged to contribute to red-teaming and other tests of the AI@UN system.

13. Sustainability: Realisation of SDGs and Agenda 2030 through three main approaches: firstly, ensuring that SDGs receive higher weights in developing AI models and tools; secondly, the AI systems themselves should be sustainable through, for example, sharing the code, building resources, and providing proper documentations and development trails; thirdly, AI solutions should be developed and deployed with environmental sustainability in mind.

14. Capacity: By developing an AI system, the UN should develop its own and wider AI capacities. Capacity development should be: (a) holistic, involving the UN Secretariat, representatives of member states, and other communities involved in UN activities; and (b) comprehensive, covering a wide range of AI capacities from a basic understanding of AI to high-end technical skills. 

15. Future-proofing: Planning and deploying systems dealing with future technological trends. Experience and expertise gathered around AI@UN should be used to deal with other emerging technologies, such as augmented/virtual reality and quantum computing. 

Opportunities in crises. AI transformation will inevitably trigger tensions due to its impact on deeper layers of how the UN functions. Likely opposition based on human fear and attachments to the status quo should be openly addressed and reframed around opportunities that AI transformation will open on individual and institutional levels. 

For instance, AI can help small and developing countries to participate in more informed and impactful ways in the work of the UN. AI can help compensate for the smaller size of their diplomatic missions and services, which must follow the same diplomatic dynamics as larger systems. An emphasis on AI will reduce current AI asymmetry.

AI can also help the UN Secretariat to refocus time and resources and spend less time on traditional paperwork, like preparing reports, to allow more work on the ground in member states where their help is critical.

Next steps. Embarking on this journey towards integrating AI into the UN’s operations is not merely a step but a leap into the future – one that demands boldness, a cooperative spirit, and an unwavering dedication to the ideals that have anchored the UN since its inception. The potential for AI to bolster the UN’s mission to uphold global peace, advance development, and champion human rights is immense. In fact, the need to adopt an open-source AI framework exceeds the need for technological innovation. The UN will be able to evolve, take the lead, and remain relevant in a rapidly changing global landscape by adopting an open approach to AI.

By leveraging the transformative power of AI, the UN can turn a looming challenge into a watershed moment, ensuring the organisation’s relevance and leadership in charting the course of human progress for all.

This text was adapted from AI@UN: Navigating the tightrope between innovation and impartiality, first published on Diplo’s blogroll.

AIatUN 1
www.diplomacy.edu

The UN faces the challenge of integrating AI in a way that maintains its impartiality and credibility, advocating for an open-source AI platform contributed to by countries, companies, and citizens to ensure transparency, inclusivity, and adherence to its core principles.


How AI chatbots master language: Insights from Saussure’s linguistics

Linguistics, intertwined with modern technology, prompts questions about how chatbots function and respond articulately to diverse inputs. Chatbots, powered by large language models (LLMs) like ChatGPT, acquire digital cognition and articulate responses using principles rooted in the linguistic theories of Ferdinand de Saussure.

Saussure’s early 20th-century work laid the groundwork for understanding language through syntax and semantics. Syntax refers to the rules governing the arrangement of words to form meaningful sentences. Saussure saw syntax as a system of conventions within a language community, interlinked with other linguistic elements like semantics. Semantics involves the study of meaning in language. Saussure introduced the concept of the sign, consisting of the signifier (sound/image) and the signified (concept), which is crucial for understanding how LLMs process and interpret word meanings.

Two humanoid robots drawn in cubism style talk as though in conversation.

How LLMs process language. LLMs like ChatGPT process and understand language through several core mechanisms:

  1. Training on vast amounts of textual data from the internet to predict the next word in a sequence
  2. Tokenisation to divide the text into smaller units
  3. Learning relationships between words and phrases for semantic understanding
  4. Using vector representations to recognise similarities and generate contextually relevant responses
  5. Leveraging transformer architecture to efficiently process long contexts and complex linguistic structures

LLMs transform text into tokenised units (signifiers) and map these to embeddings that capture their meanings (signified). The model learns these embeddings by processing vast amounts of text, identifying patterns and relationships analogous to Saussure’s linguistic structures.

Semantics and syntax in LLMs. Understanding and generating text in LLMs involves both semantic and syntactic processing. 

LLMs process semantics through (a) contextual word embeddings that capture word meanings in different contexts based on usage, (b) an attention mechanism to prioritise important words, and (c) layered contextual understanding that handles words that have multiple related meanings (polysemy) and different words with the same meaning (synonymy). The model is pre-trained on general language patterns and fine-tuned on specific datasets for enhanced semantic comprehension. 

For syntax, LLMs use (a) positional encoding to understand word order, (b) attention mechanisms to maintain sentence structure,   (c) layered processing to build complex sentences, and (d) learn probabilistic grammar from large amounts of text. Tokenisation and sequence modelling help track relationships between words, and the transformer model integrates both sentence structure and meaning at each stage, ensuring responses are both meaningful and grammatically correct. Training on diverse datasets further enhances its ability to generalise across various ways of using language, making the chatbot a powerful natural language processing tool.

Integrating Saussure’s linguistic theories with the cognitive mechanisms of large language models illuminates the inner workings of contemporary AI systems and also reinforces the enduring relevance of classical linguistic theories in the age of AI.

This text was adapted from In the beginning was the word, and the word was with the chatbot, and the word was the chatbot, first published on the Digital Watch Observatory.

generate and image of digital letters and algorithms to depict how chatbots acquire knowledge respectively how they learn by being fed with words and sentences
dig.watch

Given the profound importance of language and its various disciplines in technological developments, it is crucial to consider how chatbots function as products of advanced technology. Specifically, it contributes to understanding how chatbots learn through algorithmic cognition and how they effectively and accurately respond to diverse user queries reflecting their systems in linguistics studies



Social media giants win in free speech showdown at US Supreme Court

Social media platforms play an imperative role in people’s lives, not only in communication but also in receiving and disseminating information. At the same time, some risks may come with social media content, such the possibility of as hate speech, the spread of mis- and disinformation, and harassment. This has raised questions on the liability of social media platforms in regulating such content, as well as the role of governments in taking action. Do social media platforms have free speech rights? Can governments implement policies against social media platforms and their own content policies? The US Supreme Court tackled those questions in its decision in Moody vs. NetChoice and NetChoice, LLC vs. Paxton

NetChoice and the Computer and Communications Industry Association (CCIA), a coalition of social media companies and internet platforms, challenged the laws of two US states, Florida and Texas. These laws were enacted in 2021 amidst growing Republican party criticism of social media companies’ enforcement of their own policies

NetChoice and CCIA claimed that the Florida and Texas laws violate private companies’ first amendment rights and that governments should not be allowed to intervene in private companies’ speech policies. A group of political scientists filed an amicus brief stating that these two laws do not set a threshold as to what they consider to be hate speech and what dangerous and violent election-related speech could prevent social media platforms from moderating threats against election officials. On the other hand, officials from Texas and Florida argue that these laws aim to regulate the liability of social media platforms rather than restrict speech online while stressing that the first amendment does not apply to private businesses. One US federal appeals court invalidated Florida’s statute, while another upheld the Texas law. However, both laws were suspended pending the US Supreme Court’s final decision.

The hand of a black-robed figure holds a gavel striking its wooden base, on a desk, with the scales of justice in the background.

The supreme court decided that the lower courts’ decisions were inadequate for free speech rights under the first amendment and that the two laws are unconstitutional. In concluding its decision, the supreme court found that social media platforms are protected by the first amendment when they create content. The supreme court has also ruled that presenting a curated collection of others’ speech counts as expressive activity. 

Essentially, this sets a precedent for setting free speech rights under the first amendment for social media platforms and private businesses in the USA. Namely, US states cannot implement policies restricting their ability to regulate the content disseminated on their platforms. This could prevent governments from enacting laws leading to social media platforms losing their independence in regulating their content.


Governments steam forward with digital antitrust oversight 

In 1996 John Perry Barlow penned ‘A Declaration of the Independence of Cyberspace’. This anthology document, which reflected the libertarian internet culture of the time, was a push-back against governmental intervention and regulation of the blooming technology sector. Accordingly, governments around the world adopted a hands-off approach, under the assumption that regulation could stifle innovation. 

Almost three decades later, this understanding has radically changed. In recent years, reports published by several organisations, such as the World Bank, the Internet Society, and UNCTAD have shown a growing concentration of wealth and power in the digital economy. Data divides are particularly relevant in this context, as they lead to concentration upstream, in data-intensive technology sectors, such as AI. Against this backdrop, investigations into the potentially anti-competitive behaviour adopted by tech companies are proliferating.

A human hand holds a magnifying glass over four blocks. The first, third, and fourth blocks have green checkmarks on them. The second has a red triangle with an exclamation point inside of it.

In the EU, recent investigations have led to the first charge brought by the European Commission against a tech company under the Digital Markets Act (DMA), a law designed to curb Big Tech’s dominance and foster fair competition. According to the preliminary findings of an investigation launched in March, Apple would be in violation of the DMA. Apple’s App Store allegedly squeezes out rival marketplaces by making it more difficult for users to download apps from alternative stores, and by not allowing app developers to communicate freely and conclude contracts with their end users. Apple has been given the opportunity to review the preliminary findings, and it can still avoid a fine if it presents a satisfactory proposal to address the problem. 

Other countries are also hardening their laws on competition. The ‘Brussels effect’ and the influence of the DMA can be seen in the Digital Competition Bill, proposed by the government of India to complement existing antitrust laws. Similarly to the DMA, the law would target large companies and could introduce similarly heavy fines. In particular, tech giants would be prohibited from exploiting non-public user data and from favouring their own products or services on their platforms. They would also be barred from restricting users’ ability to download, install, or use third-party apps.

The bill is raising concern among tech companies. A US lobbying group has opposed the move, fearing its impact on business. Inspired by the common belief that dominated the tech sector in the 1990s, technology companies claim that India’s bill could stifle innovation. However, the claim seems unlikely to prosper. 

Concerns about tech sector competition are rising in the USA, traditionally an advocate for minimal regulation. The USA is tightening AI industry controls, with the DOJ and the Federal Trade Commission (FTC) dividing oversight: the FTC will regulate OpenAI and Microsoft, while the DOJ oversees Nvidia. Although less active than the EU in antitrust regulation, the US closely monitors mergers and acquisitions. This recent agreement between the two governmental bodies paved the way for antitrust investigations to be launched.

Competition is increasingly becoming a playing field with significant ​governmental activity and oversight. As countries re-assert their jurisdiction, claims of cyber independence seem a distant echo from the past.


DW Weekly #167 – 5 July 2024

 Page, Text

Dear readers, 

Welcome to another issue of the Digital Watch weekly! 

‘We have won the war on floppy disks on June 28!’ Japan’s Minister for Digital Transformation, Taro Kono cheerfully announced a few days ago. That sentence may make you think ‘Good riddance!’ or ‘Oh wow, I haven’t used one of those in ages’ (if you’re not Japanese). Or it might make you think: ‘What’s a floppy disk?’

A floppy disk, or just floppy, is a flexible removable magnetic disk (typically encased in a plastic envelope or a hard plastic shell) for storing data. Here’s the kicker: floppies can only hold between 800 KB and 2.8 MB of data, with 1.44 MB being the standard. 

Until recently, they were very much used in Japanese admin. However, two years ago, Kono declared war on floppies. He vowed that the digital agency where he is in charge will change over 1,000 government procedures that require the use of floppies so that online services can be used instead. So they did it, and they ‘won the war’.

But why were floppies in use in Japan until recently? Various explanations have been posited, from the conservative nature of bureaucracy to the floppy’s reliability. Not only does physical media offer a higher degree of authenticity of information, but floppies almost never break or lose data. Maybe it can all be summed up as: If it ain’t broke don’t fix it.

The next frontier? Fax machines, Kono said at a conference in June

Drawing of Japan’s Minister for Digital Transformation Taro Kono standing in front of a large red circle, smiling and holding up a paper decree in front of multiple microphones. At the side is a yellow garbage container overflowing with floppy disks.

In other news, a World Intellectual Property Organization (WIPO) report shows China leads in AI patents, though US firms dominate in advanced AI systems. In another major event, 10 billion passwords were leaked in the largest publication of a compilation of passwords to date.

Andrijana and the Digital Watch team


Highlights from the week of 28 June-5 July 2024

united nations headquarters in new york city

The SDOs argue that these proposals promote centralisation, which they believe would harm the internet and global economies and societies.

flag of the united nations

The non-binding resolution, initiated by China and co-sponsored by several other countries, seeks to foster international cooperation and urge organisations and financial institutions to support this cause.

abstract technology ai computing

Over the past decade, 54,000 GenAI inventions were recorded, with a significant surge occurring in the last year. China contributed over 38,200 patents, compared to nearly 6,300 from the USA.

ai brain intelligent ai technology digital graphic design electronics ai machine learning robot human brain science artificial intelligence technology innovation futuristic

The EU is far behind its 2030 AI targets, with only 11% of firms using AI. Optimistic leaders call for more investments, cooperation, and a completed Digital Single Market.

flag of japan

AI integration in Japan’s defence aims at combat speed and operational efficiency.

Apple

Apple plans to integrate ChatGPT into its devices, aiming to boost AI capabilities and meet consumer demand.

gpu chipset workforce

The programme is part of a broader $39 billion effort to enhance US chipmaking capabilities and reduce reliance on foreign suppliers.

chinese flag 1752046 1280

The initiative, which includes considerations for ethics and safety, aims to unify Chinese research efforts and position the country as a leader in international BCI standards.

massive password leak

The RockYou2024 data will be included in Cybernews’ Leaked Password Checker to help users identify if their credentials were compromised.

croatia flag close up

Authorities have been notified, and a criminal investigation is underway.

woman using laptop wood desk with cyberattack warning screen cyber security concept

The incident compromised the personal data of customers, including sensitive information such as Social Security Numbers and medical records, with unauthorised activity traced back to late October 2023.

law and justice in united states of america

The justices unanimously overturned lower court decisions due to inadequate consideration of First Amendment implications and directed further analysis.

japan banknotes

Despite the rise in cashless transactions, cash remains significant, and the redesign aims to stimulate consumer spending and productivity amid inflationary pressures.



ICYMI

What is digital immortality? And (how) can it be achieved? Watch to find out!


Upcoming

8N6ikeWA un flag blue

The UN OEWG 2021-2025 8th substantive session will focus on adopting the group’s annual progress report (APR), taking stock of the group’s discussions on threats to information security, rules, norms, and principles of responsible behaviour of states, international law, confidence-building measures (CBMs), capacity building efforts, and a regular open-ended institutional dialogue under the auspices of the UN.

hlpf

The High-level Political Forum on Sustainable Development (HLPF) 2024 will take place from 8 to 17 July under the theme ‘Reinforcing the 2030 Agenda and eradicating poverty in times of multiple crises: the effective delivery of sustainable, resilient and innovative solutions’. 

DW Weekly #166 – 28 June 2024

 Page, Text

Dear readers, 

Welcome to another issue of the Digital Watch weekly! 

In 2022, a team of Japanese scientists invented a dermis or skin equivalent, a living layer composed of cells and an extracellular matrix, in an attempt to cover the robots’ outer layer to make them look more human. Now, they are pushing forward with another invention: a way to wrap that dermis equivalent seamlessly onto the rigid surfaces of a robot.

These perforation-type anchors were inspired by the structure of skin ligaments. The researchers constructed a robotic face covered with dermis equivalent and a silicone layer. They then used perforation-type anchors to move the silicone around the corners of the robot’s mouth and, ultimately, to make a robot smile.

 Baby, Person, Face, Head, Furniture, Piggy Bank, Marie Laurencin
Image credit: Kawai et al., Perforation-type anchors inspired by skin ligament for robotic face covered with living skin, Cell Reports Physical Science

The skin equivalent’s primary advantage, researchers say, is that it is capable of self-healing. It can regenerate missing skin through cellular proliferation without any triggers. Researchers tested this previously, in 2022, on a robotic finger but have not tested it since the invention of perforation-type anchors. They noted that ‘the facial shape, with its intricate unevenness, differs significantly from the simpler, convex shape of a finger’. However, the researchers acknowledged that is a future challenge.

The fun part? The research might be of use to humans in the future: Researchers will study how the face moves to ensure the skin equivalent is applied with sufficient thickness over a robot’s face. As forming expressions such as smiles is closely linked to the development of wrinkles, researchers expect that the knowledge they gain about wrinkle formation could find applications in the cosmetics and orthopaedic surgery industries.

During this process, researchers controlled mechanical actuators beneath the dermis equivalent. The scientists note that replacing mechanical actuators with cultured muscle tissue could advance the understanding of emotions and aid treatments like facial paralysis surgery.

We can expect that while it looks quite unsettling now, the skin-equivalent-covered robot is bound to look better, more realistic, and even more akin to humans.

 Book, Comics, Publication, Person, Adult, Female, Woman, Face, Head

Monotheistic religions such as Christianity and Islam hold that God made man in his image. Christianity, in particular, holds that God is a benevolent creator who loves his creations. One might ask: Are humans determined to play God with robots?

Andrijana and the Digital Watch team


Highlights from the week of 21-28 June 2024

apple 1839363 1280

Criticism also arose over Apple’s delay in launching AI-powered features in the EU, citing DMA compliance issues.

building 1011876 1280

Microsoft has indicated its willingness to collaborate with EU regulators to find solutions.

wikileaks

US authorities have pursued Assange since 2010 following WikiLeaks’ significant disclosure of confidential files.

closeup male hand holding smartphone with blank screen background blurred flag india

India is set to mandate USB-C as the standard charging port for smartphones and tablets from June 2025, aligning with the EU’s efforts to enhance user convenience and reduce electronic waste.

flag of indonesia 1

Hackers using a variant of LockBit 3.0 ransomware have disrupted Indonesia’s national data centre, affecting immigration checks and over 200 public services, with the government refusing to pay the $8 million ransom demand.

eu flags in front of european commission

Tech companies like WhatsApp, privacy advocates, legal and security experts, and numerous EU lawmakers argue the plan threatens privacy and encryption, potentially leading to mass surveillance.

flag of usa and china on cracked concrete wall background

The regulations initially target China, Macau, and Hong Kong, with possible expansion to other regions and include various exceptions to address national interest and existing commitments.

Screenshot 2023 08 09 at 12.57.10

The company expands its AI chip offering to the Middle East through a new partnership with Qatar’s Ooredoo amid US export restrictions.

circuit board and ai micro processor artificial intelligence of digital human 3d render

Manufacturing will be outsourced to Taiwan’s TSMC, although production is expected to start later this year.

flag india computer keyboard technology concept

The decline signifies a lack of organic demand, with most current transactions stemming from banks disbursing benefits to employees.



Upcoming

UN Cybercrime Convention 1920x1080px 2

The event is aimed at identifying expectations (if any) from the concluding session of the Ad Hoc Committee to Elaborate a Comprehensive International Convention on Countering the Use of Information and Communications Technologies for Criminal Purposes. It will highlight the elements that define a positive outcome of this process, based on the insights of different experts and stakeholders.

DW Weekly #165 – 21 June 2024

 Page, Text

Dear readers, 

Welcome to another issue of the Digital Watch weekly! 

Chinese scientists created an AI military commander that mimics human military leaders in various ways, including experience, thought patterns, and personality traits, even incorporating their flaws. The catch is that this AI military commander is strictly confined to a laboratory at the Joint Operations College of the National Defence University in Shijiazhuang, Hebei and used only for virtual war games.

In these simulations, the AI is given significant control, acting as the main decision maker. This allows the People’s Liberation Army (PLA) to conduct extensive and varied war games far beyond what would be possible with the limited availability of senior human commanders. The AI’s knowledge base is capped to simulate human forgetfulness. As its memory reaches its limit, older knowledge is discarded, ensuring that only the most relevant information is retained for decision-making.

News coverage claims that AI is forbidden to lead the armed forces in China. This seems to allude to China’s military principle: ‘The Party commands the gun.’ What would stop the Party from commanding AI to command the gun in battle?
International discussions about the use of AI in the military tend to revolve around (lethal) autonomous weapon systems (LAWS) – think missiles, drones, submarines – able to make their own decisions on the battlefield. The GGE on LAWS, for instance, has been discussing issues related to human responsibility for decisions on the use of weapons systems and human-machine interaction. However, the use of AI to completely replace military commanders is, as of yet, an uncharted area.

Caricature drawing shows six identical soldiers standing in a present arms position in front of a Chinese flag. They are facing a laptop computer which has a chat bubble with the text: '>attention; >present arms; >at ease; > ... .

AI was also prevalent on the agenda of the Group of Seven (G7) summit in Italy, where Pope Francis made an unprecedented appearance, stressing the importance of human control over AI, particularly in life-or-death decisions, and calling for a ban on autonomous weapons. He called for ‘algor-ethics’, a set of global and pluralistic ethical principles, to guide AI development globally.

The G7 Leaders’ Communiqué, published last Friday, also tackled AI. The G7 leaders emphasised the need for a unified approach to manage AI’s benefits and risks. A new action plan to effectively harness AI in the workplace was announced. A credential that can be used to identify organisations supporting the implementation of the Hiroshima Process International Code of Conduct for Organizations Developing Advanced AI Systems is also in the works. Key initiatives include advancing the Hiroshima AI Process, supporting SMEs, and fostering digital inclusion and lifelong learning.

The G7 outlined a strategy to combat cyber threats, focusing on responsible state behaviour, cybersecurity enhancement across sectors, deterrence tools, and building cyber capacities through initiatives like the Cybersecurity Working Group and the Ise-Shima Cyber Group. The G7 reaffirmed support for the UN Program of Action (PoA) on ICTs as the future format for regular institutional dialogue on cybersecurity.

Additionally, the G7 established a Semiconductors G7 Point of Contact Group to coordinate semiconductor industry issues.

Another regional gathering that caught attention was EuroDIG 2024, the European version of the Internet Governance Forum (IGF), held on 17–19 June in Vilnius, Lithuania. If you didn’t follow the discussions in real-time, we’ve got you covered with our hybrid reports

And just yesterday, the Republic of Korea convened a high-level Security Council open debate on cybersecurity as part of its presidency of the Security Council. The debate centred on how the Security Council could address the harmful use of ICT in order to maintain international peace and security.

Andrijana and the Digital Watch team


Highlights from the week of 14 -21 June 2024

africa ai digital continent only the african continent with artificial intelligence digitalised

The AI Strategy aims to harness AI for development, ensuring ethical use while minimising risks and maximising opportunities.

broadband commission

The report aims to guide policymakers on AI advancements to ensure equitable digital access.

IMF

AI’s impact on jobs could widen economic inequality, affecting both white-collar and blue-collar professions.

DALL%C2%B7E 2023 11 22 22.33.01 A photo realistic image representing a conceptual conflict in semiconductor technology between China and the United States. The image features a large

Washington aims to add 11 more Chinese chipmaking factories to a restricted list and extend controls on additional equipment.

TikTok

The company claims the divestiture is technologically, commercially, and legally unfeasible and that the law violates free speech rights while unfairly targeting TikTok. However, the US Justice Department defends the legislation, still citing national security concerns.

american law symbol with hammer and us flag

Kaspersky’s proposed mitigating measures were deemed insufficient, and violations of the new rules will result in fines or criminal charges.

robber wearing black hoodie against digitally generated russian national flag

The ongoing ICC investigation could potentially set a legal precedent for cyberwarfare.

hacker working code network

The breach, initially blamed on a third-party contractor and a misconfigured firewall, was later traced to an IT service desk operator at Medibank who inadvertently provided a hacker access to the system.

hacker 3342696 1280

The incident underscores the increasing cyber threats facing the healthcare industry and the vulnerabilities within its infrastructure.

google 959059 1280

Alphabet’s Google faces a complaint from Austrian group NOYB for allegedly tracking Chrome users without proper consent, despite promoting privacy safeguards.

senegal

To bolster Senegal’s digital transformation, Senegal Numérique SA has entered a partnership with the African Digital Development Agency (ADD) to share best practices and enhance the interoperability of government information systems and services.



ICYMI

EuroDIG 2024 banner

EuroDIG 2024 was held in Vilnius, Lithuania, from June 17–19. If you couldn’t make it to the conference, don’t fret. DiploAI-generated reports are ready for you to dive into.


Reading corner

generate and image of digital letters and algorithms to depict how chatbots acquire knowledge respectively how they learn by being fed with words and sentences

Given the profound importance of language and its various disciplines in technological developments, it is crucial to consider how chatbots function as products of advanced technology. Specifically, it contributes to understanding how chatbots learn through algorithmic cognition and how they effectively and accurately respond to diverse user queries reflecting their systems in linguistics studies.

g7

The Leaders of the G7 issued a communiqué after their summit in Apulia, Italy, tackling AI, chips and cybersecurity. Read the communiqué in full.

G7 Italy 2024

Pope Francis spoke during the second day of the G7 summit in Borgo Egnazia, Italy, on 14 June, focusing on the ethical implications and potential dangers of AI. Read his speech in full.