Analysis Archives | Page 2 of 2 | Digital Watch Observatory

5G Transformation: The power of good policy

The global rollout of 5G networks has been met with considerable excitement, and rightly so. While the promise of faster data speeds has captured much of the spotlight, the true transformational potential of 5G extends far beyond mere internet speed enhancements. Across continents, from the bustling metropolises of North America to the vibrant landscapes of Africa, a diverse array of strategies and approaches is shaping the future of 5G transformation connectivity. As policymakers grapple with the intricacies of crafting effective 5G spectrum policies, it’s essential to understand how these policies are intrinsically linked to achieving the wider benefits of this groundbreaking technology.

The spectrum: A valuable resource

At the heart of 5G technology is the radio spectrum, a finite and valuable resource allocated by governments to mobile network operators. These spectrum bands determine the speed, coverage, and reliability of wireless networks. In 2023, there’s a high demand for mid-band and the millimeter-wave spectrum, both essential for delivering the anticipated 5G transformation.

City, Chart, Diagram, Plan, Plot, Urban, Neighborhood, Metropolis — Frequency bands of 5G networks [picture from digi.com]

Policy imperatives to ensure low latency

Ultra-low latency is one of 5G’s defining features, enabling real-time communication and interaction over the internet. Policy decisions that prioritise and allocate specific spectrum bands for applications that require low latency, such as telemedicine and autonomous vehicles, can have a profound impact on their effectiveness and safety. Policymakers must prioritise the allocation of spectrum for latency-sensitive applications while also accommodating the growing data demands of traditional mobile services.

The US Federal Communications Commission (FCC) launched its 5G FAST Plan in 2018. This initiative facilitates the deployment of 5G infrastructure by streamlining regulations and accelerating spectrum availability. As part of the programme, the FCC conducted auctions for spectrum bands suitable for 5G, such as the 24 GHz and 28 GHz bands, to support high-frequency, low-latency applications.
The EU introduced the 5G Action Plan in 2016 as part of its broader Digital Single Market strategy. The plan emphasises cooperation among EU member states to create the conditions needed for 5G deployment, including favourable spectrum policies.
China launched its National 5G Strategy in 2019, outlining a comprehensive roadmap for 5G development. The strategy includes policies to allocate and optimise spectrum resources for 5G networks.The Independent Communications Authority of South Africa (ICASA) is actively exploring spectrum policies to accommodate 5G. ICASA has published draft regulations for the use of high-demand spectrum, including the 3.5 GHz and 2.6 GHz bands, which are crucial for 5G deployment. ICASA’s efforts to regulate spectrum have been praised by the Wi-Fi Alliance for their role in advancing Wi-Fi technology and connectivity in Africa. ICASA aims to amend radio frequency regulations to stimulate digital development, investment, and innovation in the telecom sector for public benefit.

Enabling massive Internet of Things connectivity

The International Telecommunication Union (ITU) has classified 5G mobile network services in three categories: Enhanced Mobile Broadband (eMBB), Ultra-reliable and Low-latency Communications (uRLLC), and Massive Machine Type Communications (mMTC).The mMTC service was created specifically to enable an enormous volume of small data packets to be collected from large numbers of devices simultaneously; this is the case with internet of things (IoT applications. mMTC classified 5G as the first network designed for Internet of Things from the ground up.

City, Metropolis, Urban, Architecture, Building, Cityscape, 5g speeds, Neighborhood, Office Building, Lighting, Outdoors — 5g communication network is important for the Internet of Things powered ‘Smart Cities’

The IoT stands as a cornerstone of 5G transformation potential; 5G is expected to unleash a massive 5G IoT ecosystem where networks can serve the communication needs for billions of connected devices, with the appropriate trade-offs between speed, latency, and cost. However, this potential hinges on the availability of sufficient spectrum for the massive device connectivity that the IoT needs. The demands that the IoT places on cellular networks vary by application, often requiring remote device management. And as connectivity and speed (especially even very short network dropouts) are mission critical for remotely-operated devices, URLLC and 5G Massive MIMO radio access technologies offer key ingredients for effective IoT operations.

Effective 5G spectrum policies must allocate dedicated bands for IoT devices while ensuring interference-free communication. Standards in Releases 14 and 15 of the Third Generation Partnership Project (3GPP) the solve most of the commercial bottlenecks to facilitate the vision of 5G and the huge IoT market.

Diverse approaches to spectrum allocation

The USA’s spectrum allocation strategy is centered around auctions as its primary methodology. The FCC has been at the forefront of this approach, conducting auctions for various frequency bands. This auction-driven strategy allows network operators to bid for licenses, enabling them to gain access to specific frequency ranges. Notably, the focus has been on making the mid-band spectrum available, with a significant emphasis on cybersecurity.

Its proactive stance has marked South Korea’s approach to spectrum allocation. Among the pioneers in launching commercial 5G services, the South Korean government facilitated early spectrum auctions. As a result, they allocated critical frequency bands, such as 3.5 GHz and 28 GHz, for 5G deployment. This forward-looking strategy not only contributed to the rapid adoption of 5G within the nation, but also positioned South Korea as a global leader in the 5G revolution.

The Korea Fair Trade Commission (KFTC), South Korea’s antitrust regulator, has fined three domestic mobile carriers a total of 33.6 billion won ($25.06 million) for exaggerating 5G speeds. [link]

The EU champions spectrum harmonisation to enable seamless cross-border connectivity. The identification of the 26 GHz band for 5G in the Radio Spectrum Policy Group (RSPG) decision further supports the development of a coordinated approach. By aligning policies across member states, the EU aims to eliminate fragmentation and ensure a cohesive 5G experience.

Moreover, many African countries are in the process of identifying and allocating spectrum for 5G deployment. Governments and regulatory bodies have considered various frequency bands, such as the C-Band (around 3.5 GHz) and the millimeter-wave bands (above 24 GHz), for 5G services. Some African nations have issued trial licenses to telecommunications operators to conduct 5G trials and test deployments. Thesehelp operators understand the technical challenges and opportunities associated with 5G in the African context. For example, in South Africa, ICASA is developing a framework for 5G spectrum allocation. Their approach encompasses license conditions, coverage requirements, and the possibility of sharing spectrum resources.

Kenya is in the process of exploring opportunities to release additional spectrum to facilitate 5G deployment. The Communications Authority of Kenya is contemplating reallocating the 700 and 800 MHz bands for mobile broadband use, including 5G services.

A well-structured spectrum management framework serves as the guiding principle for equitable and efficient allocation of this resource. These frameworks include regulatory approaches like exclusive licensing, lightly-managed sharing, and license-exempt usage. Sharing frameworks enable coexistence, from simple co-primary sharing to multi-tiered arrangements. Static sharing uses techniques such as FDMA and CDMA, while Dynamic Spectrum Sharing (DSS) allows users to access spectrum as needed.

In conclusion, the intricate world of 5G spectrum policies profoundly shapes the path of 5G’s transformative journey. Beyond speed enhancements, global strategies spotlighted here reveal the interplay of technology and governance.

From South Korea’s spectrum leadership to the EU’s harmonisation and Africa’s context-specific solutions to challenges, each of these approaches underscores the link between policies and 5G’s potential. These efforts are indispensable to foster optimal policies for future development.

Today’s decisions will echo into the future, moulding 5G’s global impact. This intricate interweaving emphasises 5G’s capabilities and policy’s role in driving unprecedented connectivity, innovation, and societal change.

Internet shutdowns: Can we find solutions?

By Bojana Kovač

Internet shutdowns present intentional disruptions of the internet or of electronic communications, which can occur nationwide or in specific locations. They can be partial or in the form of total internet blackouts, blocking people from using the internet in its entirety. According to a research conducted by Shurfshark, 4.24 billion individuals have been affected, globally, in the first half of 2023, where 82 internet restrictions affected 29 countries.

It has been estimated that Iran imposed the largest number of internet restrictions, while India proved to be a world leader in internet shutdowns in 2022. While the EU member states have not experienced total or partial blackouts, the statement made on France Info a few months ago by the EU’s Internal Market Commissioner, Thierry Breton, during the riots in France, raised a few eyebrows. Breton suggested that online platforms could be suspended for failing to promptly remove illegal content, especially in case of riots and violent protests.

x0aGqhqa9Gh0eFNwos7dhDMHK09UyXtMhDa5cyRDVJ56BBXolPRXHdIqp4YoCWkEfJTpEF2XpbI7DTQcUwrV6Zlt2FJp oqap6HvEorsFT5xj5tQ6gaPa5QawfvanHQFK256 ekS8oT9KdkL4ehVSto — Author: Pietro Naj-Oleari

This announcement led to more than 60 civil rights NGO’s seeking clarification, because they were concerned that the Digital Services Act (DSA) might be misused as a means of censorship. Brenton then clarified, in his comments, that the temporary suspension should be the last resort in case platforms fail to remove illegal content. These extreme circumstances include systemic failures in addressing infringements related to calls for violence or manslaughter. Significantly, Breton underscored that the courts will make the final decision on such actions, ensuring a fair and impartial review process.

On a global scale, Shurfshark found that since 2015, there have been:

107 disruptions in Africa,
585 disruptions in Asia,
15 disruptions in Europe,
9 disruptions in North America, and
42 disruptions in South America.

Shutdowns being used as a tool for repressing fundamental human rights

Most countries, if not all, justify shutdowns as a means of maintaining national security or for the prevention of false information, among others. However, the internet shutdowns have, in some cases, become a tool for digital authoritarianism with derogatory effects on human rights, including the right to free speech, access to information, freedom of assembly, and development.

The UN Special Rapporteur on the rights to freedom of peaceful assembly and association, Clément Nyaletsossi Voule, stated that the internet shutdowns violate international human rights law and cannot be justified. On a regional level, the European Court of Human Rights (ECtHR) ruled in the Cengiz and Others v. Türkiye case that ‘the Internet plays an important role in enhancing the public’s access to news and facilitating the dissemination of information in general.’ The case concerned the decision of a Turkish court to block access to Google Sites of an internet site owner who faced criminal proceedings for insulting the memory of Atatürk. While this was not a total blackout case, the ECtHR ruled that even a limited effect on internet restriction constitutes a violation of the freedom of expression.

The African Commission on Human and Peoples’ Rights (ACHPR) also condemned the imposing of internet shutdowns in its 2016 report, and urged African states to ensure effective protection of internet rights.

The case of Gabon

The most recent case of an internet blackout was the one in Gabon, which occurred on 26 August 2023, the day the presidential and legislative elections took place. Minister Rodrigue Mboumba Bissawou, announced the internet blackout and a nightly 7:00 pm – 6:00 am curfew from Sunday on state television.

Bissawou claimed that the blackout is aimed at preventing the spread of false information and countering the spread of violence. On 30 August 2023, following the military coup, the independent and non-partisan internet monitor Netblocks reported the gradual restoration of internet connectivity. It should be noted that this is not the first time the country has faced an internet blackout, in view of the same thing occurring in 2019 during the attempted coup.

Electrical Device — Internet shutdowns: Can we find solutions?

Responsibility of the ISPs and telecoms

Taking into account the regional and international instruments, fundamental human rights are being infringed. Considering such restrictions as imposed by governmental authorities, the UN Human Rights Council, in its 2022 annual report, called on private companies, including Internet Service Providers (ISPs) and telecoms, to explore all lawful measures to challenge the implementation of internet shutdowns. Namely, it called on them to ‘carry out adequate human rights due diligence to identify, prevent, mitigate, and assess the risks of ordered Internet shutdowns when they enter and leave markets.’

In 2022, the Human Rights & Human Rights Resource Center urged telecommunications companies to take action to ensure human rights protection due to the growing uncertainty of internet access worldwide. The report also highlighted the companies’ responsibilities under the UN Guiding Principles on Business and Human Rights to adopt human rights policies to

uphold the rights of users or customers,
enhance transparency on requests by governments for internet shutdowns,
negotiate human rights-compliant licensing agreements,
adopt effective and efficient remedial processes.

The problem, however, lies in the fact that most of the telecoms are state-owned and controlled by current governments. Even in cases of foreign ISPs or telecoms, there is a high chance that they would sell their services to the governments because they would have to comply with national laws, which would violate human rights laws under the jurisdiction in which the companies are based. An example of this would be Telenor, a company which operated its services in Myanmar while the country adopted the draft cybersecurity bill, allowing the military junta to order internet shutdowns. Despite the Norwegian telecom company’s opposition to Myanmar’s draft cybersecurity bill for failing to ensure effective human rights protection, Telenor complied with the military’s requests and sold its operations in Myanmar. This raised many concerns among civil society, as Telenor was accused of not being transparent when selling its services. Digital civil rights organisation Access Now criticised Telenor’s lack of transparency, accusing it of making it even more difficult to develop mitigation strategies to avoid serious human rights abuses.

Is there a way out?

Internet shutdowns have intensified over the years, and urgent actions are necessary to prevent further human rights violations. It is evident that the governments are unable or unwilling to take any action, while the private actors have not yet been established at the level of being able to guarantee internet access during disruption. Therefore, until the governments take a more robust action to ensure internet access and end human rights violations, users should be educated on how to prepare themselves, expecting a shutdown. Access Now recommends downloading several Virtual Private Networks (VPNs) in advance if there is a risk of an internet shutdown, while governments often resort to blocking access to VPN providers. At the same time, the privacy policy of each VPN shall be checked beforehand, as not all VPNs guarantee effective privacy protection.

The consequences of Meta’s multilingual content moderation strategies

By Alicia Shepherd-Vega

About 8 million Ethiopians use the world’s most popular social media platform, Facebook, daily. Its use, of course, is confined to the parameters of their specific speech communities. In Ethiopia, there are some 86 languages spoken by the population of 120.3 million), but 2 (Amharic and Oromo) are spoken by two-thirds of the population. Amharic is the second most popular language.

Like most countries across the globe, the use of social media in Ethiopia is ubiquitous. What sets Ethiopia apart, though, as with many countries in the Global South, are the issues that arise with developments designed by the Global North for the Global North context. This perspective becomes apparent when one views social media usage from the angle of linguistics.

Content moderation and at-risk countries (ARCs)

Increased social media usage has recently engendered a proliferation of policy responses, particularly concerning content moderation. The situation is no different in Ethiopia. Increasingly, Ethiopians blame Meta and other tech giants for the rate and range within which conflict spreads across the country. For instance, Meta faces a lawsuit filed by the son of an Ethiopian academic, Mareg Amare, whose father was assassinated in November 2021. The lawsuit claims that Meta failed to delete life-threatening posts from the platform, categorised as hate speech against Mareg’s father. Meta, earlier, had assured the global public that a wide variety of context-sensitive strategies, tactics, and tools were used to moderate content on its platform. The strategies for this and other such promises was never published, until the leak of the so-called Facebook files, brought to the fore results of key studies conducted by Meta, such as the harmful effects experienced by users of Meta’s platforms, Facebook and Instagram.

Meta employees have also complained of human rights violations, including overexposure to traumatic content, including abuse, human trafficking, ethnic violence, organ selling, and pornography, without a safety net of employee mental health benefits. Earlier this year, workers at Sama, a subsidiary of Meta in Kenya, received a ruling from a local court that Meta must reinstate their jobs after dismissing them for complaints about working under these conditions and attempts to unionise. The court later ruled that the company is also responsible for their mental health, given their overexposure to violent content on the job.

The disparity in the application of content moderation strategies, tactics, and tools used by the tech giant is also a matter of concern. Crosscheck or XCheck, a quality control measure used by Facebook for high-profile accounts, for example, shields millions of VIPs, such as government officials, from the enforcement of established content moderation rules; on the flip side, inadequate safeguards on the platform have coincided with attacks on political dissidents. Hate speech is said to increase by some 300% amidst bloody riots. This is no surprise, given Facebook’s permissiveness in the sharing and recycling of fake news and plagiarised and radical content.

In the case of Ethiopia, the platform has catalysed conflict. In October 2021, Dejene Assefa, a political activist with over 120 million followers, called for supporters to pick up arms against the Tigrayan ethnic group. The post was shared about 900 times and received 2,000 reactions before it was taken down. During this period, it was reported that the federal army had also waged war against the Tigrayans because of an attack on its forces. Calls for an attack against the group proliferated on the platform, many of which were linked to violent occurrences. According to a former Google data scientist, the situation was reminiscent of what occurred in Rwanda in 1994. In another case, the deaths of 150 persons and the arrest of 2000 others coincided with the protests that ensued following the assassination of activist Hachalu Hundessa after he had campaigned on Facebook for better treatment of the Oromo ethnic group. The incident led to a further increase in hate speech on the platform, including from several diasporic groups. Consequently, Facebook translated its community standards into Amharic and Oromo for the first time.

In light of ongoing conflicts in Ethiopia, Facebook labelled the country a first tier ‘at risk country’, among others like the USA, India, and Brazil. ARCs are at risk of platform discourse inciting offline violence. As a safeguard, war rooms are usually set up to monitor network activities in these countries. For developing countries like Ethiopia, such privileges are not extended by Facebook. In fact, although the Facebook platform can facilitate 110 languages, it only can review 70. At the end of 2021, Ethiopia had no misinformation or hate speech classifiers and had the lowest completion rate for user reports on the platform. User reports help Meta identify problematic content. The problem here was that the interfaces used for such reports lacked local language support.

Languages are only added when a situation becomes openly and obviously untenable, as was the case in Ethiopia. It usually takes Facebook at least one year to introduce the most basic automated tools. By 2022, amidst the outcry for better moderation in Ethiopia, Facebook partnered with local moderation companies PesaCheck and AFP Fact Check and began moderating content in the two languages; however, only five persons were deployed to scan content posted by the 7 million Ethiopian users. Facebook principally uses automation for analysing content in Ethiopia.

AI and low-resource languages

AI tools are principally used for automatic content moderation. The company claims Generative AI in the form of Large Language Models (LLMs) is the most scalable and best suited for network-based systems like Facebook. These LLMs are developed via natural language processing (NLP), which allows the models to read and write texts like humans do. According to Meta, whether a model is trained in one or more languages, such as XLM-R and Few-Shot Learner, they are used to moderate over 90% of content on its platform, including content in languages on which the models have not been trained.

These LLMs train on enormous amounts of data from one or more languages. They identify patterns from higher-resourced languages in a process termed cross-lingual transfer, and apply these patterns to lower-resourced languages, to identify and process harmful content. Languages with a resource gap are languages that do not have high-quality digitised data available to train models. However, one challenge with monolingual and multilingual models is that they have consistently missed the mark on analysing violent content appropriately in English. The situation has been worse for other languages, particularly in the case of low-resource languages like Amharic and other Ethiopian languages.

AI models and network-based systems have the following limitations :

They rely on machine-translated texts, which sometimes contain errors and lack nuance.
Network effects are complex for developers, so it is sometimes difficult to identify, diagnose, or fix the problem when models fail.
They cannot produce the same quality of work in all languages. One size does not fit all.
They fail to account for the psycho-social context of local-language speakers, especially in high-risk situations.
They cannot parse the peculiarities of a lingua franca and apply them to specific dialects.
Machine language (ML) models depend on previously-seen features, which makes them easy to evade as humans can couch meaning in various forms.
NLP tools require clear, consistent definitions of the type of speech to be identified. This is difficult to ascertain from policy debates around content moderation and social media mining.
ML models reflect the bias in their training data.
The highest-performing models accessible today only achieve between 70%-75% accuracy rates, meaning one in every four posts will likely be treated inaccurately. Accuracy in ML is also subjective, as the measurement varies from developer to developer.
ML tools used to make subjective predictions, like whether someone might become radicalised, can be impossible to validate.

According to Natasha Duarte and Emma Llansó of the Centre of Democracy and Technology,

Today’s tools for automating social media content analysis have limited ability to parse the nuanced meaning of human communication, or to detect the intent or motivation of the speaker… without proper safeguards these tools can facilitate overboard censorship and a biased enforcement of laws and of platforms’ terms of service.

In essence, given that existing LLM models are proven to be ineffective in analysing human language on Facebook, should tech giants like Facebook be allowed to enforce platform policies around their use for content moderation, there is a risk of stymying free speech as well as the leakage of these ill-informed policies into national and international legal frameworks. According to Duarte and Llansó, this may lead to human rights and liberties violations.

Human languages and hate speech detection

The use and spread of hate speech are taken seriously by UN countries, as evidenced by General Assembly resolution A/res/59/309. Effective analysis of human language requires that fundamental tenets responsible for language formation and use be considered. Except for some African languages not yet thoroughly studied, most human languages are categorised into six main families: Indo-European, from which we have European languages like English and Spanish, and North American, South American, and some Asian languages. The other categories are Sino-Tibetan, Niger-Congo, Afro-Asiatic, Austronesian and Trans-New Guinea. The Ethiopian languages Oromo, Somali, and Afar fall within the Cushitic and Omotic subcategories of the Afro-Asiatic family, whereas Amharic falls within the Semitic subgroup of that family.

This primary level of linguistic distinction is crucial to understanding the differences in language patterns, be they phonemic, phonetic, morphological, syntactic or semantic. These variations, however, are minimal when compared with the variations brought about by social context, mood, tone, audience, demographics, and environmental factors, to name a few. Analysing human language in an online setting like Facebook becomes particularly complex, given its mainly text-based nature and the moderator’s inability to observe non-linguistic cues.

Variations in language are even more complex in the case of hate speech, given the role played by factors like intense emotions. Davidson et al. (2017) describe hate speech as ‘speech that targets disadvantaged social groups in a manner that is potentially harmful to them, … and in a way that can promote violence or social disorder’. It intends to be derogatory, humiliate or insult. To add to the complexity, hate speech and extremism are also often difficult to distinguish from other types of speech, such as political activism and news reporting. Hate speech can also be mistaken for offensive words. And offensive words can be used in non-offensive contexts such as music lyrics, taunting or gaming. Other factors such as gender, audience, ethnicity and race also play a vital role in deciphering the meaning behind language.

On the level of dialectology, parlance, such as slang, can be used as offensive language or hate speech, depending partly on whether it is directed at someone or not. For instance, ‘life’s a bi*ch’ is considered offensive language for some models, but it can be considered hate speech when directed at a person. Yet, hate speech does not always contain offensive words. Consider the words of Dejene Assefa in the case mentioned above, ‘the war is with those you grew up with, your neighbour… If you can rid your forest of these thorns… victory will be yours’. Slurs also, whether offensive or not, can emit hate. ‘They are foreign filth’ (containing non-offensive wording used for hate speech) and ‘White people need those weapons to defend themselves from the subhuman trash these spicks unleash on us’ provide examples. Overall, hate speech reflects our subjective biases. For instance, people tend to label racist and homophobic language as hate speech but sexist language as merely offensive. This also has implications for analysing language accurately. Who is the analyst? And in terms of models, whose data was the model trained on?

The complexities mentioned above are further compounded when translating or interpreting between languages. The probability of transliteration (translating words on their phonemic level) increases with machine-enabled translations such as Google Translate. With translations, misunderstanding grows across language families, particularly when one language does not contain the vocabulary, characters, conceptions, or cultural traits associated with the other language, an occurrence referred to by machine-learning engineers as the UNK problem.

Yet, from all indications, Facebook and other tech giants will invariably continue to experiment with using one LLM to moderate all languages on their platforms. For instance, this year, Google announced that its new speech model will encompass the world’s 1000 most spoken languages. Innovators are also trying to develop models to bridge the gap between human language and LLMs. Lesan, a Berlin-based startup, built the first general machine translation service for Tigrinya. It partners with Tigrinya-speaking communities to scan texts and build custom character recognition tools, which can turn the texts into machine-readable forms. The company also partnered with the Distributed AI Research Institute (DAIR) to develop an open-source tool for identifying languages spoken in Ethiopia and detecting harmful speech in them.

Conclusion

In cases like that of Ethiopia, it is best first to understand the broader system and paradigm at play. The situation is typical of the pull and push typical of a globalised world where changes in the developed world wittingly or unwittingly create a pull on the rest of the world, drawing them into spaces where they subsequently realise they do not fit. It is from the consequent discomfort that the push emerges. What is now evident is that the developers of the technology and the powers that sanctioned its use globally did not anticipate the peculiarities of this use case. Unfortunately, this is not atypical of an industry that embraces agility as a modus operandi.

It is, therefore, more critical now than ever that international mechanisms and frameworks, including a multistakeholder, cross-disciplinary approach to decision-making, be inculcated in public and private sector technological innovations at the local level, particularly in the case of rapidly scalable solutions emerging from the Global North. It is also essential that tech giants be held responsible for the equitable distribution within and across countries with the resources needed for the optimal implementation of safety protocols concerning content moderation. To this end, it would serve Facebook and other tech giants well to partner with startups like Lesan.

It is imperative that a sufficient quantity of qualified persons with on-the-job mental health benefits be engaged. to deal with the specific issue of analysing human languages, which still have innumerable unknowns and unknown unknowns. The use of AI and network-based systems can only be as effective as the humans behind the technologies and processes. Moreover, Facebook users will continue to adapt their use of language. It is reckless to anticipate that these models would be able to adjust to or predict all human adaptive strategies. And even if these models eventually can do so, the present and interim impact, as seen in Ethiopia and other countries, is far too costly in human rights and lives.

Finally, linguistics, like all disciplines and languages, is still evolving. It is irresponsible, therefore, to pin any, let alone all, languages down to one model without the foreknowledge of dire consequences.

Networked journalism and online media: Reimagining trust for digital reporters

By Luka Avramović

Shifting landscapes and multiple discourses

The digital changes in the topography of journalism have, for better or worse, resulted from two essential shifts in how information circulates in society. One tectonic change shows in the amount and types of actors that engage in news reporting have massively increased due to the accessibility afforded by the Fourth Industrial (or Digital) Revolution. From non-governmental private companies, such as social media conglomerates Meta and Alphabet, to individuals with information power, like Julian Assange and Elon Musk, to everyday consumers like you and me, the new engagement paradigm is in full swing. For another, and leading on from the first, the sheer abundance of information being shared through different media and platforms has reinforced the plurality of discourse. Different sources, communities, filter bubbles, and political or personal biases shape what information appears, who is writing it, and where and why.

A question of (anti-)social media?

In a rapidly evolving technological landscape, the internet and social media have revolutionised how information is disseminated. However, this transformation does not necessarily translate to improved journalism.

With greater accessibility and connectivity for both citizens and reporters, concerns are mounting over the proliferation of biassed information. Recent pivotal examples are the COVID-19 pandemic, ongoing conflicts in Ukraine or in central Africa and elections worldwide (particularly in the USA and Türkiye – where the interplay between governments, private media companies, and individuals has increased political tensions across society). The surge in online news outlets and social media has further exacerbated the situation, providing a platform for individuals to disseminate biassed, misleading, or inaccurate information. Consequently, this will reach a wider audience, giving rise to a host of emerging issues in the media landscape, like disengagement or polarisation.

Recent trends in news consumption

Over the past year, trust in the news has experienced a notable decline, dropping by an additional 2 percentage points across various markets worldwide, according to Reuters 2023 Digital News Report. In many countries, this setback has undone the progress made during the peak of the COVID-19 pandemic, when trust in broadcast and paper news sources had witnessed an upswing.

Their research has also shown that users on rapidly popularising short-form platforms such as TikTok, Instagram, and Snapchat are notably more inclined to pay attention to updates from celebrities, influencers, and social media personalities rather than relying on professional journalists. In stark contrast, and counterintuitively, Facebook and Twitter maintain their status as platforms where news media, journalists and reporters remain central in shaping conversations.

Three improvements any journalist should institute

In the quest to tackle trust issues between sources and journalists, the crux of the matter lies in the power balance. The powers at play are the selection of a plurality of sources or one absolute ‘truth’. In other words, do we recognise that there is no universal truth and seek to include diverse perspectives, or do we trust that there is a ubiquitous, truth that can be proved and reported on. Acknowledging this fundamental challenge, experts worldwide emphasise that journalists must take charge and adapt to the digital era. By embracing the online environment to their advantage, journalists can consolidate their position and credibility, thereby enhancing public trust in their work.

To achieve this end, journalists must think of information no longer as a product, but as a service that the news media should be responsible for delivering. With the abundance of interconnected information sources in today’s society, expecting reporters to single-handedly provide all-encompassing news coverage has become impractical. Instead, experts propose a shift in focus, emphasising journalists’ role as arbiters, curators, and information filters (Broersma & Graham, 2016; Beckett, 2018; Dahlgren, 2009). By becoming gatekeepers of trustworthy information, they can guide audiences through the sea of media confusion that characterises modern life.

Second, journalists might want to foster open discourse by providing contrarian opinions while removing themselves from the perceived role of absolute authority. Doing this, allows journalists or reporters to effectively and accessibly communicate knowledge while allowing room for healthy debate and critical examination.

Third, the dynamic power of so-called citizen journalists should not be underestimated. Journalists should see the increased involvement of members of the public gathering and spreading news and information as a tool, not as a constraint. At a time when many news organisations face staff cutbacks, citizen journalists have emerged as valuable contributors who play a crucial role in monitoring society using online resources and social media.

Takeaway

Amidst the ongoing mis- and dis- information crises, credible sources and information filtering emerge as a potent antidote, fostering a fresh perspective on information management within the field of news journalism. Good information and good journalism empower people through knowledge and allow individuals to make informed decisions. Emphasising the pivotal role of reliable news reporting, this approach bolsters the belief that trustworthy journalism remains integral to the fabric of society.

References

Beckett, C. (2018). The Paradox of Power for Journalism: Back to the Future of News [new book]. London School of Economics. https://blogs.lse.ac.uk/polis/2018/11/23/the-paradox-of-power-for-journalism-back-to-the-future-of-news-new-book/

Broersma, M. & Graham, T. (2016). ‘Chapter 6: Tipping the Balance of Power: Social Media and the Transformation of Political Journalism’, in Burns, A. (ed.) The Routledge Companion to Social Media and Politics. New York: Routledge, pp. 89–103.

Dahlgren, P. (2009). Media and Political Engagement: Citizens, Communication and Democracy. Cambridge; New York: Cambridge University Press, pp. 172–181.

Newman, N., Fletcher, R., Eddy, K., Robertson, C. T., & Kleis Nielsen, R. (2023). Reuters Institute Digital News Report 2023. Reuters Institute for the Study of Journalism. https://reutersinstitute.politics.ox.ac.uk/sites/default/files/2023-06/Digital_News_Report_2023.pdf

Worldcoin: Eye-scanning ID is here

Worldcoin history

Back in the golden era of blockchain (2018-2019), when questions raised in everyday conversation were promised to be solved by this technology, a group of people started working on an ambitious project called Worldcoin. This project tried to find a solution for the challenge of unique online identification (our so-called digital identity). In particular, Worldcoin developed a system for recording and storing users’ digital biometric data and offering them a reward in the form of digital tokens. The data that Worldcoin gathered were iris scans. To join the user base, people would go to the designated location and consent to have their irises scanned. This was done using a shiny spherical object they named Orb. In the short period that Orb collected data, a significant database of human irises was collected. Least-developed countries had the most users, as was generally expected, because Worldcoin guaranteed tokens as incentives (i.e. money) ‘simply for being human’.

Wordcoin’s Orb custom biometric device. Source: worldcoin.org

The technology behind the identification scheme is the following: Iris scans were digitally obfuscated using a hashing function (this is a cryptography technique in which one set of digital data can be encrypted to match a unique digital key for reading these data). That unique hash was added to the database as each person’s unique identifier. Even though this data is encrypted, significant concerns were raised that a possible data breach could create a privacy and data nightmare. The crypto community had serious concerns about a scary dystopian future, undermining the project. The Worldcoin project was almost forgotten and was considered one of the most ambitious and yet obscure in the crypto community.

The rebirth and rebranding of Worldcoin

Fast forward to 2022, when the Worldcoin project leader, Sam Altman, became globally famous as OpenAI’s CEO. Only half a year after the ambitious ChatGPT launch and global excitement about the predictive language models, Altman pushed the ‘old’ Worldcoin idea into the public space again.

Earlier this week, the ‘new’ Worldcoin project launched worldwide, but with one significant difference. It is being publicised as ‘a new identity and financial network owned by everyone’. The rebranding is important, because now, the project team claims that what they are building is not distinguishable from the Public Key Infrastructure (PKI) deployed by big companies or the technical internet society. PKI is a set of standards, software, and hardware used in digital certificates and for managing public-key encryption. This is done via certificate authorities, with one of the most notable implementations being the HTTPS protocol used for secure web browsing. Worldcoin will use a cryptographic technique known as zero-knowledge proof or ZKP.

This obfuscating technique allows verification that the ‘given statement is true while avoiding conveying any information to the verifier beyond the mere fact of the statement’s truth’. This technique is used in some privacy-oriented cryptocurrencies, and it demonstrates the possibility of user-defined online privacy divisions allowing options to decide what information you want to share with whom. For example, your browser doesn’t need to know all your credentials and data. In fact, it only uses your IP (for geolocation) and information like gender or age for advertising or other purposes. ZKP solutions were tested in COVID-19 tracking apps and are at the core of the EU’s new Digital Identity proposal. Significant concerns exist about the gatekeepers of certificate authorities that store the data. This issue is crucial for sensitive data, such as biometric data collected by the Orbs.

How is this data stored? Is any unencrypted version of the iris data stored in a secure manner (e.g. in the Orb’s temporary internal memory)? Who has access to this data? Or even worse, can it end up on the black market or be misused somehow? In its launch report, Worldcoin claimed that: ‘The Orb sets a high bar to defend against scalable attacks; however, no hardware system interacting with the physical world can achieve perfect security’.

One way of looking at Worldcoin is that it is very similar to Apple’s PKI, and there is nothing to be worried about. One difference with Worldcoin is that part of the identifier data will be stored inside Ethereum’s public, open-source blockchain, while World IDs are issued on the Worldcoin protocol The Worldcoin protocol was developed by Tools For Humanity, a company established by the founders of the original Worldcoin project: Alex Blania and Sam Altman. The design ensures that no trusted third party can introduce risks of data handling or accountability related to it. Users have control of the process. However, the past has shown us that human users are usually the weakest link. Human factors include the very real risk that users will share their biometric data like they share their ultrasounds. So far technology has not found a way to limit voluntary violations of privacy and security. The UK data watchdogs at the Information Commissioner’s Office, have already announced a probe into Worldcoin’s privacy and data protection practices.

Interactive infographic available at https://worldcoin.org/home shows a global map marked with different colors of dots showing global users of World IDs, transactions, Activities, and Milestones. — The Worldcoin home page shows this interactive map of its global users.

Another part of the project also makes it significantly different from known PKI schemes, and it’s a digital currency reward that actors get for sharing their biometric data.

Worldcoin was not accessible in the USA at its launch, and anyone wishing to participate had to confirm that they were outside the USA. The Worldcoin launch report clearly stated that tokens distributed in the system will be only available where laws allow this to happen.

Why is this important?

Aside from the technological, privacy and data protection, and other ethical questions raised, the financial incentives and infrastructure that are underlying the project will also be scrutinised.

Only a couple of years ago, Meta (then Facebook) and Mark Zuckerberg announced the launch of the Libra digital token, which, in their words, could offer a solution for cross-globe payments in different currencies across all Meta apps (Facebook, Instagram, and WhatsApp). Meta signed agreements with major payment institutions like Visa and Mastercard and giant online retailers like Ebay, but US legislators torpedoed the project. In three separate hearings in front of the US regulators in the Senate and House, the USA made it clear that no digital coin issued by a private company can be considered an international means of payment, particularly if it is pegged to or in any way related to the US dollar, which is regarded as a global reserve currency. The Libra project was shut down after two years, and mentions of Libra were erased from company websites.

Digital currencies issued by private companies remain of primary interest to major state powers and international financial organisations, like the USA and the UK and the Bank for International Settlements or the G7’s Financial Stability Board. This, in fact, might be a more significant obstacle for Worldcoin than data collection and privacy issues.

Worldcoin promotes the ’proof of personhood’ idea, which establishes an individual as both human and unique, and might become indispensable to discern and identify AI identities, like bots, bot farms, and ‘fake humans’. We will certainly hear more about this project.

MOVEit hack: what is it and why is it important?

A string of disclosures

On 31 May, Progress Software Corporation disclosed that its managed file transfer (MFT) software, MOVEit Transfer, is susceptible to a critical SQL injection vulnerability, which allows unauthenticated attackers to acquire access to MOVEit Transfer databases.

On 2 June, the vulnerability received the designation CVE-2023-34362. CVE stands for Common Vulnerabilities and Exposures ID number, which is assigned for publicly disclosed vulnerabilities. Once a CVE is assigned, vendors, industry and cybersecurity researchers can exchange information to develop remediation.

On 9 July, Progress announced additional vulnerabilities (CVE-2023-35036), which were identified during code reviews. The company also released a patch for new vulnerabilities. On 15 June, a 3rd vulnerability was announced (CVE-2023-35708).

Threat actors have attacked more than 162 known victims, including the BBC, Ofcom, British Airways, Ernst and Young, Siemens Energy, Schneider Electric, UCLA, AbbVie, and several government agencies with these zero-day vulnerabilities. Sources also report the compromise of the personal data of more than 15.5 million individuals.

Behind the attack

Microsoft attributed the MOVEit hack to Lace Tempest, a threat actor known for ransomware attacks and for running the extortion website of the CLOP ransomware group, data theft, and extortion attacks. On 6 June, the CLOP ransomware gang posted a communication to their leak site demanding that victims contact them before 14 June to negotiate extortion fees for deleting stolen data.

The identity and whereabouts of the CLOP gang remain unknown to the public. However, security researchers believe the group is either linked to Russia or comprises Russian-speaking individuals.

Supply chain security flaws

The MOVEit hack has again highlighted that supply chain security is a significant concern for industries and the public sector. Across the supply chains, who is responsible for what? And how can we ensure cross-sectoral and cross-border cooperation between multiple actors that mitigate security risks?

While national cybersecurity agencies continue publishing guidance on mapping and securing supply chains, the industry implements good practices for reducing vulnerabilities and building secure ICT infrastructures. Still, organisations have different levels of maturity and resources to respond effectively. Luckily, there are ongoing discussions at different levels to address these topics: from international levels to advance the implementation of the relevant UN GGE norms to reduce vulnerabilities and secure supply chains, such as the Geneva Dialogue, to national and industry-specific discussions to develop and adopt new security measures (e.g. SBOM).

Another challenge lies in conducting effective investigations, with the participation of several states and/or private partners, to identify a threat actor and stop the activity.

Digital policy trends in June 2023

Governing AI: What are the appropriate AI guardrails?

AI governance remains the number one trend in digital policy as national, regional and global efforts to shape AI guardrails continue.

The EU’s risk-based approach

The European Parliament’s approval of the AI Act is a groundbreaking development. This regulation classifies AI systems based on risk levels and safeguards of civil rights, with severe fines for violations. Next in the legislative process is the so-called trialogues, where the European Parliament, the EU Council, and the Commission have to agree on a final version of the act; there are expectations that this agreement will be reached by the end of the year.

A new study from Stanford suggests that leading AI models are still far off of the responsible AI standards set by the AI Act (the version agreed in the EP), notably lacking transparency on risk mitigation measures. But some in the industry argue that the rules impose too heavy a regulatory burden. A recent open letter signed by some of the largest European companies (e.g. Airbus, Renault, Siemens) notes that the AI Act could harm the EU’s competitiveness and could compel them to move out of the EU to less restrictive jurisdictions. Companies are, in fact, doing their best to shape things: For example, OpenAI lobbied successfully in the EU that the forthcoming AI Act should not consider OpenAI’s general-purpose AI systems to be high risk, which would trigger stringent legal requirements like transparency, traceability, and human oversight. OpenAI’s arguments align with those previously employed by the lobbying efforts of Microsoft and Google, which argued that stringent regulation should be imposed only on companies that explicitly apply AI to high-risk use cases, not on companies that build general-purpose AI systems.

Given the EU’s track record on data protection rules, its proposed AI Act was anticipated to serve as an inspiration to other jurisdictions. In June, Chile’s Parliament initiated discussions on a proposed AI Bill, focusing on legal and ethical aspects of AI’s development, distribution, commercialisation, and use.

More regional rules are in the works: It has been revealed that ASEAN countries are planning an AI guide that will tackle governance and ethics. In particular, it will address the use of AI for generating misinformation online. The guide is expected to be adopted in 2024. Strong dynamism will occur during Singapore’s chairmanship of ASEAN in 2024.

Business-friendlier approaches

Considering that Singapore itself is taking a collaborative approach to AI governance and is focused on working with businesses to promote responsible AI practices, the ASEAN guide is not likely to be particularly stringent (watch out, EU?). Softer, more collaborative approaches are also expected to be formulated in Japan and the UK, which believe such an approach will help them position themselves as AI leaders.

Another country that is taking a more collaborative approach to AI governance is the USA. Last month, President Biden met with Big Tech critics from civil society to discuss AI’s potential risks and implications of AI on democracy, including the dissemination of misinformation and the exacerbation of political polarisation. The US Commerce Department will create a public working group to address the potential benefits and risks of generative AI and develop guidelines to effectively manage those risks. The working group will be led by NIST and comprise representatives from various sectors, including industry, academia, and government.

Patchwork

As countries continue their AI race, we might end up with a patchwork of legislation, rules and guidelines that might espouse conflicting values and priorities. It is no surprise that calls for global rules and an international body are also gaining traction. A future global AI agency inspired by the International Atomic Energy Agency (IAEA), an idea first put forward by OpenAI CEO Sam Altman, has garnered support from UN Secretary-General Antonio Guterres.

France is advocating for global AI regulation, with President Macron proposing that the G7 and the Organisation for Economic Co-operation and Development (OECD) would be good platforms for this purpose. France wants to work alongside the EU’s AI Act while advocating for global regulations and also intends to collaborate with the USA in developing rules and guidelines for AI. Similarly, Microsoft’s President Brad Smith called for collaboration between the EU, the USA, and G7 nations, adding India and Indonesia to the list, to establish AI governance based on shared values and principles.

In plain sight: SDGs as guardrails

However, the road to global regulations is typically long and politically tricky. Its success is not guaranteed either. Diplo’s Executive Director Dr Jovan Kurbalija argues that humanity is missing valuable AI guardrails that are in plain sight: the SDGs. They are current, comprehensive, strong, stringently researched, and immediately applicable. They already have global legitimacy and are not centralised and imposing. These are just a handful of reasons why the SDGs can play a crucial role; there are 15 reasons why we should use SDGs for governing AI.

Digital identification schemes gain traction

Actors worldwide are pushing for more robust, secure and inclusive digital ID systems and underlying policies.

Businessman using fingerprint identification to access and protecting personal information data

The OECD Council approved a new set of recommendations on the governance of digital identity centred on three pillars. The first addresses the need for systems to be user-centred and integrated with existing non-digital systems. The second focuses on strengthening the governance structure of the existing digital systems to address security and privacy concerns, while the third pillar addresses the cross-border use of digital identity.

Most recently, the EU Parliament and the Council reached a preliminary agreement on the main aspects of the digital identity framework put forward by the Commission in 2021. Previously, several EU financial institutions cautioned that specific sections of the regulation are open to interpretation and could require significant investments by the financial sector, merchants, and global acceptance networks.

At the national level, a number of countries have adopted regulatory and policy frameworks for digital identification. Australia released the National Strategy for Identity Resilience to promote trust in the identity system across the country, while Bhutan endorsed the proposed National Digital Identity Bill, except for two clauses that await deliberation in the joint sitting of the Parliament. The Sri Lanka Unique Digital Identity Project (SL-UDI) is underway, and the Thai government introduced the ThaID mobile app to simplify access to services requiring identity confirmation.

Content moderation: gearing up for the DSA

Preparations for the DSA are in full swing, even though the European Commission has already faced its first legal challenge over the DSA, and it did not come from Big Tech as many would have expected. German e-commerce company Zalando filed a lawsuit against the Commission, contesting the categorisation of Zalando as a systemic, very large platform and criticising the lack of transparency and consistency in platform designation under the DSA. Zalando argues that it does not meet the requirements for such classification and does not present the same systemic risks as Big Tech.

Meanwhile, European Commissioner for Internal Market Thierry Breton visited Big Tech executives in Silicon Valley to remind them of their obligations under the DSA. Although Twitter owner Musk previously said that Twitter would comply with the DSA content moderation rules, Breton visited the company headquarters to perform a stress test to evaluate Twitter’s handling of potentially problematic tweets as defined by EU regulators. Breton also visited the CEOs of Meta, OpenAI, and Nvidia. Meta agreed to a stress test in July to assess the EU’s online content regulations, the decision prompted by Breton’s call for immediate action by Meta regarding its content targeting children.

People, Person, Crowd, Adult, Male, Man, Face, Head, Audience, Lecture, Indoors, Room, Seminar, Speech, Thierry Breton — European Commissioner for Internal Market Thierry Breton. Credit: European Commission

The potential of the EU to exert its political and legal power over Big Tech will be demonstrated in the coming months, with the DSA becoming fully applicable in early 2024.

ChatGPT and GDPR: Balancing AI innovation with data protection

By Feodora Hamza

OpenAI’s ChatGPT has gained widespread attention for its ability to generate human-like text when responding to prompts. However, after months of celebration for OpenAI and ChatGPT, the company is now facing legal action from several European data protection authorities who believe that it has scraped people’s personal data, without their consent. The Italian Data Protection Authority has temporarily blocked the use of ChatGPT as a precautionary measure, while French, German, Irish, and Canadian data regulators are also investigating how OpenAI collects and uses data. In addition, the European Data Protection Board set up an EU-wide task force to coordinate investigations and enforcement concerning ChatGPT, leading to a heated discussion on the use of AI language models and raising important ethical and regulatory issues, particularly those involving data protection and privacy.

Concerns around GDPR compliance: How can generative AI comply with data protection rules such as GDPR?

According to Italian authorities, OpenAI’s disclosure regarding its collection of user data during the post-training phase of its system, specifically chat logs of interactions with ChatGPT, is not entirely transparent. This raises concerns about compliance with General Data Protection Regulation (GDPR) provisions that aim to safeguard the privacy and personal data of EU citizens, such as the principles of transparency, purpose limitation, data minimisation, and data subject rights.

As a condition for lifting the ban it imposed on ChatGPT, Italy has outlined the steps OpenAI must take. These steps include obtaining user consent for data scraping or demonstrating a legitimate interest in collecting the data, which is established when a company processes personal data within a client relationship, for direct marketing purposes, to prevent fraudulent activities, or to safeguard the network and information security of its IT systems. In addition, the company must provide users with an explanation of how ChatGPT utilises their data and offer them the option to have their data erased, or refuse permission for the program to use it.

Electronics, Hardware, Computer Hardware — Padlock symbol for computer data protection system. Source: Envato Elements

Steps towards GDPR compliance: OpenAI’s updated privacy policy and opt-out feature

OpenAI has updated its privacy policy, describing its practices for gathering, utilising, and safeguarding personal data. In a GPT-4 technical paper, the company stated that publicly available personal information may be included in the training data and that OpenAI endeavours to ensure people’s privacy by incorporating models to eliminate personal data from training data ’where feasible’. In addition, OpenAI allows now for an incognito mode on ChatGPT to enhance its GDPR compliance efforts, safeguard users’ privacy, and prevent the storage of personal information, granting users greater control over the use of their data.

The company’s choice to offer an opt-out feature comes amid mounting pressure from European data protection regulators concerning the firm’s data collection and usage practices. Italy has demanded OpenAI’s compliance with the GDPR by April 30. In response, OpenAI implemented a user opt-out form and the ability to object to personal data being used in ChatGPT, allowing Italy to restore access to the platform in the country. This move is a positive step towards empowering individuals to manage their data.

Challenges in deleting inaccurate or unwanted information from AI systems remain

However, the issue of deleting inaccurate or unwanted information from AI systems in compliance with GDPR is more challenging. Although some companies have been instructed to delete algorithms developed from unauthorised data, eliminating all personal data used to train models remains challenging. The problem arises because machine learning models often have complex black box architectures that make it difficult to understand how a given data point or set of data points is being used. As a result, models often have to be retrained with a smaller dataset in order to exclude specific data, which is time-consuming and costly for companies.

Data protection experts argue that the OpenAI could have saved itself a lot of trouble by building in robust data record-keeping from the start. Instead, it is common in the AI industry to build data sets for AI models by scraping the web indiscriminately and then outsourcing the work of removing duplicates or irrelevant data points, filtering unwanted things, and fixing typos. In AI development, the dominant paradigm is that the more training data – the better. OpenAI’s GPT-3 model was trained on a massive 570 GB of data. These methods, and the sheer size of the data set, mean that tech companies tend to not have full understanding of what has gone into training their models.

While many criticise the GDPR for being unexciting and hampering innovation, experts argue that the legislation serves as a model for companies to improve their practices when they are compelled to comply with it. It is presently the sole means available to individuals to exercise any authority over their digital lives and data in a world that is becoming progressively automated.

The impact on the future of generative AI: The need for ongoing dialogue and collaboration between AI developers, users, and regulators

This highlights the need for ongoing dialogue and collaboration between AI developers, users, and regulators to ensure that the technology is used in a responsible and ethical manner. It seems that ChatGPT is facing a rough ride with Europe’s privacy watchdogs. The Italian ban seems to have been the beginning, since OpenAI has not set up a local headquarters in one of the EU countries yet, exposing it to further investigations and bans from any member country’s data protection authority.

However, while the EU regulators are still wrapping their head around the regulatory implications of and for generative AI, companies like OpenAI continue to benefit and monetise from the lack of regulation in this area. With the EU’s Artificial Intelligence Act being passed soon, the EU aims to address the gaps of the GDPR when regulating AI and inspire similar initiatives being proposed in other countries. It seems the impact of generative AI models on privacy will probably be on the regulators’ agenda for many years to come.

How search engines make money and why being the default search engine matters

By Kaarika Das and Arvin Kamberi

Samsung, the maker of millions of smartphones with preinstalled Google Search, is reportedly in talks to replace Google with Bing as the default search provider on its devices. This is the first instance of a threat confronting Google’s long-standing dominance over the search business. Despite Alphabet’s diversified segments, its core business and majority profit accrue from Google Search, which accounted for US$162 billion of US$279.8 billion of Alphabet’s total revenue last year. Naturally, Google’s top agenda is to protect its core business and retain its position as the default search engine in electronic devices like tablets, mobiles, or laptops.

A critical question arises about the underlying business model of online search engines like Google, Bing, Baidu, Yandex, and Yahoo. What do these search engines stand to gain by being the default devices search engine? Let us examine how search engines generate revenue while allowing users to explore the internet for information and content for free.

The profit model of search engines

Search engines make money primarily through advertising (billions of dollars yearly from its Google Ads platform). The working mechanism is as follows: Whenever users can enter a search query into a search engine, the search engine provides a list of web pages and other content related to the search query, including advertisements. Advertisers pay search engines to display sponsored results when users search for specific keywords. These ads typically appear at the top and/or bottom of Search Engine Results Pages (SERPs) and are labelled as ‘sponsored’ or ‘ad’. Search engines get paid based on the number of clicks these ads get. This model is popularly known as the PPC (Pay-Per-Click).

Apart from sponsored listing, search engines also track user data for targeted advertising, using people’s search history. Search engines can easily gather information about users’ search history, preferences, and behaviours. This is done through cookies, IP address tracking, device and browser fingerprinting, and other technologies. Search engines then use these data points to profile their users to improve the targeting of advertisements. For example, if a user frequently searches for details about recipes and food, the search engine may display advertisements for restaurants and related food ingredient products. Thus, the user search history effectively helps improve search engine algorithms and enhances search accuracy by identifying patterns in user behaviour. In capitalising on user data, search engines allow advertisers to manage their advertisements using strategies such as ad scheduling, geotargeting, and device targeting – all made possible because of accumulated user history data!

Google, magnifying glass — Google making money from search engine. Image generated by DALL-E/OpenAI.

The power of default

Let us now delve into the edge granted to a search engine by being the default setup. Regardless of the default search engine, people can always change their search engine on their respective devices based on personal preferences. Despite the absence of any exclusivity, there is massive inertia to change the default search engine. It happens because the effort required to manually navigate to a different search engine to perform search functions makes the transition process a hassle, especially for ordinary people. Parallelly, technologically challenged people may not be aware of alternative search engines and might have no explicit preference for a specific search engine. Even with awareness of alternatives, the effectiveness, performance, and security of the search engine paired with their current device remains unapproved and may lead to apprehension among users.

Therefore, a default search engine further provides a sense of security (however misleading) as its performance and device compatibility are assumed to be vetted by the manufacturers. As a result, being the default search engine is advantageous for search engines as it provides them with a broader audience base leading to increased traffic alongside greater brand recognition. Thus, being the default search engine is vital for a search engine’s success as having large traffic ensures that search engines remain attractive to advertisers, their primary source of revenue – the higher the number of search engine users, the dearer the advertising space becomes, generating better returns.

For users, however, pre-installed search engines deprive them of the choice to select their preferred alternative and select those search engines that do not track user details. In 2019, the European Commission stated that Google had an unfair advantage by pre-installing its Chrome browser and Google search app on Android smartphones and notebooks. To circumvent antitrust concerns, in early 2020, Google enabled Android smartphones and tablets sold in the European Economic Area (EEA) to show a ‘choice screen’ that offered users four search engines to choose from.

While Google pays billions to device manufacturers like Samsung and Apple to remain the default search engine, the ongoing AI leap in the industry has enormous ramifications for the future of internet search and its ensuring business model. With unprecedented developments in AI and search engine functionality integrated with AI, the tussle of search rivals battling for popularity and influence is set to continue.