Digital standards

Anthropic aims to decode AI ‘black box’ within two years​

Anthropic CEO Dario Amodei has unveiled an ambitious plan to make AI systems more transparent by 2027. In a recent essay titled ‘The Urgency of Interpretability,’ Amodei highlighted the pressing need to understand the inner workings of AI models.

He expressed concern over deploying highly autonomous systems without a clear grasp of their decision-making processes, deeming it ‘basically unacceptable’ for humanity to remain ignorant of how these systems function.

Anthropic is at the forefront of mechanistic interpretability, a field dedicated to deciphering the decision-making pathways of AI models. Despite these advancements, Amodei emphasized that much more research is needed to fully decode these complex systems.​

Looking ahead, Amodei envisions conducting ‘brain scans’ or ‘MRIs’ of advanced AI models to detect potential issues like tendencies to deceive or seek power. He believes that achieving this level of interpretability could take five to ten years but is essential for the safe deployment of future AI systems.

Amodei also called on industry peers, including OpenAI and Google DeepMind, to intensify their research efforts in this area and urged governments to implement ‘light-touch’ regulations to promote transparency and safety in AI development.​

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot! 

Meta cuts jobs in Reality Labs

Meta has announced layoffs within its Reality Labs division, impacting Oculus Studios and hardware development teams. Among those affected is the team behind Supernatural, a popular VR fitness app that Meta acquired for over $400 million.

The company stated that these restructuring efforts aim to improve efficiency and focus on developing future mixed reality experiences, particularly in fitness and gaming. Despite reaffirming its commitment to VR and mixed reality, Meta’s moves reflect its Quest headset business challenges.

While its smart glasses partnership with Ray-Ban has exceeded sales expectations, Quest devices continue to underperform, with the latest Quest 3S already seeing discounts less than a year after release.

Why does it matter?

The layoffs signal Meta’s attempt to streamline operations as it navigates a shifting market for virtual and mixed reality. Although the company promises ongoing support for its VR communities, these changes highlight the pressures Meta faces in turning its ambitious metaverse and hardware ventures into sustainable success.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Microsoft’s Surface ad uses generative AI without anyone noticing

Microsoft recently revealed that it created a minute-long ad for its Surface Pro and Surface Laptop using generative AI, but the twist is that no one seemed to notice the AI elements, even though the ad has been online for nearly three months.

Released on January 30th, the ad features a mix of real footage and AI-generated content, with some AI-generated visuals corrected and integrated with live shots.

The AI tools were first used to generate the script, storyboards, and pitch deck for the ad. From there, a combination of text prompts and sample images helped generate visuals, which were iterated on and refined with image and video generators like Hailuo and Kling.

Creative director Cisco McCarthy explained that it took thousands of prompts to achieve the desired results, although the process ultimately saved the team around 90% of the time and cost typically needed for such a production.

Despite the AI involvement, most viewers didn’t notice the difference. The ad has received over 40,000 views on YouTube, but none of the top comments suggest AI was used. The quick-cut editing style helped mask the AI output’s flaws, demonstrating how powerful generative AI has become in the right hands.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

MTN confirms cybersecurity breach and data exposure

MTN Group has confirmed a cybersecurity breach that exposed personal data of some customers in certain markets. The telecom giant assured the public, however, that its core infrastructure remains secure and fully operational.

The breach involved an unknown third party gaining unauthorised access to parts of MTN’s systems, though the company emphasised that critical services, including mobile money and digital wallets, were unaffected.

In a statement released on Thursday, MTN clarified that investigations are ongoing, but no evidence suggests any compromise of its central infrastructure, such as its network, billing, or financial service platforms.

MTN has alerted the law enforcement of South Africa and is collaborating with regulatory bodies in the affected regions.

The company urged customers to take steps to safeguard their data, such as monitoring financial statements, using strong passwords, and being cautious with suspicious communications.

MTN also recommended enabling multi-factor authentication and avoiding sharing sensitive information like PINs or passwords through unsecured channels.

While investigations continue, MTN has committed to providing updates as more details emerge, reiterating its dedication to transparency and customer protection.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

ChatGPT expands Deep Research to more users

A new feature introduced by ChatGPT in February, called Deep Research, is gradually becoming available across its user base. This includes subscribers on the Plus, Team, and Pro plans, while even those using the free ChatGPT app on iOS and Android can now access a simplified version.

Designed to carry out in-depth reports and analyses within minutes, Deep Research uses OpenAI’s o3 model to perform tasks that would otherwise take people hours to complete.

Instead of limiting access to paid users alone, OpenAI has rolled out a lightweight version powered by its o4-mini AI model for free users. Although responses are shorter, the company insists the quality and depth remain comparable.

The more efficient model also helps reduce costs, while delivering what OpenAI calls ‘nearly as intelligent’ results as the full version.

The feature’s capabilities stretch from suggesting personalised product purchases like cars or TVs, to helping with complex decisions such as choosing a university or analysing market trends.

Free-tier users are currently allowed up to five Deep Research tasks each month, whereas Plus and Team plans get ten full and fifteen lightweight tasks. Pro users enjoy a generous 125 tasks of each version per month, and EDU and Enterprise plans will begin access next week.

Once users hit their full version limit, they’ll be automatically shifted to the lightweight tool instead of losing access altogether. Meanwhile, Google’s GeminiAI offers a similar function for its paying customers, also aiming to deliver quick, human-level research and analysis.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Google stopped Motorola from using Perplexity as default assistant

A senior executive at Perplexity AI has testified that Google prevented Motorola from setting the AI startup’s assistant as the default on new smartphones, citing restrictive contracts.

The testimony came during a trial seeking remedies after Google was found to have unlawfully maintained a search monopoly.

Although Motorola will preload the app, it cannot make it the default due to binding agreements with Google. Perplexity’s Chief Business Officer, described the difficulty of replacing Google’s assistant on Android phones, saying Google’s terms create an environment where device makers fear losing revenue.

The CEO added that ongoing negotiations with other companies only became possible due to pressure from the US Department of Justice’s antitrust case.

The Justice Department is asking the court to ban Google from paying for default placements, which would also affect its AI products like Gemini.

Meanwhile, Perplexity is developing its own browser, Comet, and voiced concern about any Chrome sale undermining open-source access. The company does not support OpenAI’s interest in acquiring the browser, citing past inconsistencies in its open-source commitments.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Perplexity expands iPhone app with voice features

AI research firm Perplexity has rolled out a new voice assistant for iPhones, expanding its app’s functionality to include reminders, email writing, and third-party services like ride-booking.

The assistant allows for continuous voice interaction even when the app is running in the background, although it cannot access system-level features due to Apple’s limitations. First launched on Android in January, the AI now supports multiple apps and can play media or draft emails via default Apple apps.

Users can activate it using the Action button on newer iPhones, but some features still require manual input depending on system permissions. The assistant is free to use, with limitations on the number of messages, while a £20/month subscription lifts those restrictions.

Despite comparisons with Siri, Perplexity lacks screen or camera-sharing capabilities, though it can search content from podcasts and YouTube. Developers say the update marks a significant step towards offering an AI assistant that rivals native options.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

White House condemns EU fines on Apple and Meta

The White House has strongly criticised the EU after landmark fines were imposed on Apple and Meta Platforms, describing the penalties as a ‘novel form of economic extortion’ that the US would not tolerate.

The European Commission fined Apple €500 million and Meta €200 million under the Digital Markets Act (DMA), a new law designed to rein in the power of dominant tech giants.

Rather than viewing the DMA as a fair attempt to promote market competition, US officials called it ‘discriminatory’ and claimed it unfairly targets American firms, undermines innovation, and restricts civil liberties.

The White House warned that such extraterritorial measures would be treated as trade barriers and hinted at retaliation.

At the same time, tensions were mounting on another front, with US Treasury Secretary Scott Bessent acknowledging that tariffs between the US and China were unsustainable.

He said both sides must lower their tariffs, currently as high as 145 per cent, instead of expecting unilateral moves, suggesting a potential thaw in the ongoing trade war.

President Trump, while indicating openness to cutting Chinese import duties, also threatened to raise the existing 25 per cent tariff on Canadian car imports. He said the US should focus on building its own vehicles instead of relying on foreign manufacturers.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Microsoft expands rewards for reporting AI vulnerabilities

Microsoft has announced an expanded bug bounty initiative, offering up to $30,000 for researchers who uncover critical vulnerabilities in AI features within Dynamics 365 and the Power Platform.

The programme aims to strengthen security in enterprise software by encouraging ethical hackers to identify and report risks before cybercriminals can exploit them.

Rather than relying on general severity scales, Microsoft has introduced an AI-specific vulnerability classification system. It highlights prompt injection attacks, data poisoning during training, and techniques like model stealing and training data reconstruction that could expose sensitive information.

Highest payouts are reserved for flaws that allow attackers to access other users’ data or perform privileged actions without their consent.

The company urges researchers to use free trials of its services, such as PowerApps and AI Builder, to identify weaknesses. Detailed product documentation is provided to help participants understand the systems they are testing.

Even reports that don’t qualify for a financial reward can still lead to recognition if they result in improved defences.

The AI bounty initiative is part of Microsoft’s wider commitment to collaborative cybersecurity. With AI becoming more deeply integrated into enterprise software, the company says it is more important than ever to identify vulnerabilities early instead of waiting for security breaches to occur.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Ubisoft under fire for forcing online connection in offline games

French video game publisher Ubisoft is facing a formal privacy complaint from European advocacy group noyb for requiring players to stay online even when enjoying single-player games.

The complaint, lodged with Austria’s data protection authority, accuses Ubisoft of violating EU privacy laws by collecting personal data without consent.

Noyb argues that Ubisoft makes players connect to the internet and log into a Ubisoft account unnecessarily, even when they are not interacting with other users.

Instead of limiting data collection to essential functions, noyb claims the company contacts external servers, including Google and Amazon, over 150 times during gameplay. This, they say, reveals a broader surveillance practice hidden beneath the surface.

Ubisoft, known for blockbuster titles like Assassin’s Creed and Far Cry, has not yet explained why such data collection is needed for offline play.

The complainant who examined the traffic found that Ubisoft gathers login and browsing data and uses third-party tools, practices that, under GDPR rules, require explicit user permission. Instead of offering transparency, Ubisoft reportedly failed to justify these invasive practices.

Noyb is calling on regulators to demand deletion of all data collected without a clear legal basis and to fine Ubisoft €92 million. They argue that consumers, who already pay steep prices for video games, should not have to sacrifice their privacy in the process.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!