Artificial intelligence Archives | Page 25 of 695

Yale proposal targets transparency gap in AI development

Researchers at Yale’s Digital Ethics Center have proposed a copyleft-style licensing framework intended to increase transparency around generative AI models trained on open-source software.

The proposal, called the Contextual Copyleft AI License, would adapt principles from free and open-source software licensing to generative AI. Under the model, AI systems trained on open-source code could be treated as derivative works, requiring developers to make key information about model architecture and training data freely available.

The researchers argue that such a framework could give open-source software developers more control over how their code is used in AI development. They also say it could encourage more genuinely open AI models and reduce ‘open washing’, where systems are marketed as open despite keeping important components closed.

The proposal comes amid wider debates over AI transparency, copyright and the role of open-source software in the development of generative AI. The researchers conclude that the approach may be legally feasible under current copyright law, provided that training AI models on open-source software is not treated as fair use.

The study also notes that open generative AI models can create risks because they may be used to generate deceptive or harmful content. The researchers argue that licensing approaches need to work alongside regulatory safeguards, including rules designed to limit manipulative or deceptive uses of AI.

Why does it matter?

The proposal addresses a central transparency gap in AI development: many generative AI systems rely on open-source software but do not disclose enough about how that software is used, which data is involved, or how the resulting models work. If similar licensing approaches gained traction, they could reshape debates over AI openness, developer rights, copyright and accountability. The proposal also shows how open-source governance tools are being reconsidered for AI systems whose risks and dependencies differ from traditional software.

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

Anthropic AI restrictions reignite debate over AI sovereignty

US government restriction on foreign access to Anthropic’s Fable 5 and Mythos 5 models has triggered broader concerns about AI sovereignty among American allies. The move has raised questions about whether governments and companies outside the United States can reliably depend on frontier AI systems controlled by US firms and subject to national security restrictions.

The directive reportedly required Anthropic to prevent non-American users, including foreign nationals working inside the company, from accessing the models. Anthropic responded by suspending access more broadly, stating that this was the only practical way to comply with the directive.

The immediate dispute centres on concerns that Fable 5 could be jailbroken and used beyond its intended safeguards. However, the broader impact extends beyond one company or one model. Governments, security agencies and companies that had secured access to Anthropic’s most advanced systems reportedly saw those permissions withdrawn overnight.

The Anthropic cutoff has been particularly sensitive for US allies. Reports indicate that the restrictions extended even to partners in the Five Eyes intelligence alliance, including Australia, the U K, Canada and New Zealand. The UK’s AI Security Institute, which has played a leading role in testing and evaluating advanced models, was also reportedly affected.

The episode has strengthened arguments that countries may need greater sovereign AI capabilities rather than relying heavily on frontier models controlled by foreign providers. For allies, the question is not only whether they can access advanced AI systems today, but whether that access can be withdrawn suddenly because of US policy decisions, export controls or national security interventions.

The episode also highlights a difficult policy trade-off for the United States. The United States has a strategic lead in frontier AI and may seek to prevent the most capable systems from being misused or accessed by adversaries. Yet applying broad restrictions to allies and foreign employees could damage trust, disrupt research and push other countries to accelerate domestic AI development.

For middle powers, building AI sovereignty will not be straightforward. Training frontier models requires advanced chips, large-scale compute infrastructure, talent and capital, all of which remain concentrated in a small number of countries and firms. Restrictions on chip exports could also limit the ability of allies to build independent alternatives.

The dispute, therefore, points to a wider geopolitical shift in AI governance. As frontier AI models become more capable, access to them is increasingly being treated as a matter of national security. That could force governments to rethink procurement, cloud dependence, AI testing partnerships and long-term strategies for technological sovereignty.

Why does it matter?

The episode illustrates how access to advanced AI systems is becoming a strategic issue rather than simply a commercial service. As frontier models become increasingly important for research, cybersecurity, defence, innovation and economic competitiveness, governments are beginning to view access controls through the lens of national security and geopolitical influence.

The case also highlights a growing tension between AI leadership and international trust. While countries may seek to restrict access to powerful systems to prevent misuse, abrupt limitations affecting allies can encourage efforts to build domestic AI capabilities and reduce dependence on foreign providers. As a result, debates about AI sovereignty, technological autonomy and strategic resilience are likely to become increasingly central to digital policy worldwide.

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

Japanese researchers develop interpretable AI for materials discovery

Researchers in Japan have developed an interpretable AI method to explain how AI models make predictions in materials discovery. The method analyses features learned by a trained AI model and uses them to identify relationships between atomic structure and optical spectra.

The study was led by researchers from the Institute of Science Tokyo, in collaboration with Tohoku University. The work is expected to be published in the journal Advanced Intelligent Discovery.

AI is increasingly used in materials research to predict how materials behave based on atomic structure. Such models can accelerate materials discovery and reduce reliance on trial-and-error experimentation, but many operate as black boxes, making it difficult to understand how they arrive at specific predictions.

The researchers addressed this problem by analysing a trained AI model that predicts optical absorption spectra from atomic structural data. They extracted features from the model’s internal layers and clustered materials according to shared structural and spectral characteristics.

The team used an atomistic line graph neural network trained on data from 2,681 metal oxides, chalcogenides, and related compounds. The clustering process classified materials into groups sharing structural characteristics such as elemental composition, atomic coordination, bond lengths, bond angles and similar spectral signatures.

According to the researchers, the model learned meaningful relationships between atomic structure and material properties without being explicitly provided oxidation states or electronic configurations as input. The interpretable AI method could therefore help researchers identify the factors behind desired spectral shapes and support more rational materials design.

The approach could also be applied beyond optical absorption spectra. Researchers said the approach could also help explain how atomic arrangements influence other material properties under varying conditions, such as temperature and pressure, opening new possibilities for designing materials with targeted characteristics.

Why does it matter?

One of the main challenges facing the use of AI in scientific research is explainability. While AI systems can identify patterns and generate accurate predictions, researchers often need to understand the reasoning behind those predictions before they can confidently apply them in experimental settings.

By revealing how AI models connect atomic structures with material properties, interpretable AI could make machine learning a more effective tool for scientific discovery. The approach may help accelerate the development of advanced materials for applications ranging from renewable energy and electronics to sensors and next-generation manufacturing, while improving trust in AI-assisted research.

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

Apple delays Siri AI rollout on iOS and iPadOS in EU, citing DMA requirements

Apple has announced that its new Siri AI features will not be available to users in the European Union on iOS 27 and iPadOS 27 when the software is released later this year, citing concerns related to compliance with the EU’s Digital Markets Act (DMA).

According to the company, discussions with European regulators have not resulted in an agreement on how the new AI features could be introduced while maintaining what Apple describes as necessary privacy and security protections.

Apple said the features will remain available to EU users on macOS 27 and visionOS 27. However, users in the bloc will not have access to Siri AI on iPhone, iPad, or Apple Watch, as the watchOS functionality depends on a paired iPhone with Siri AI support.

The company stated that the DMA’s interoperability requirements would require broader access for competing virtual assistants to device functionality and user data than Apple considers appropriate from a privacy and security perspective.

Apple also said it proposed a solution called Trusted System Agent, which it described as an intermediary framework intended to provide third-party virtual assistants with access to device capabilities while maintaining additional security protections. According to the company, it also proposed a phased rollout of Siri AI in the EU while this framework was being developed.

The company said the European Commission did not accept its proposals and that there is currently no timeline for the availability of Siri AI on iOS and iPadOS in the EU.

The announcement highlights ongoing discussions between major technology companies and the EU regulators on implementing the Digital Markets Act. The DMA seeks to increase competition in digital markets by requiring designated gatekeepers to provide greater interoperability and access to certain platform services.

The European Commission has previously stated that the objective of the regulation is to promote contestability and fairness in digital markets while providing users and businesses with greater choice.

Apple’s decision means that some AI features announced at the company’s Worldwide Developers Conference (WWDC26) will not initially be available to EU users on mobile devices. These include new AI-powered assistance capabilities, expanded visual intelligence features, and AI tools integrated across iOS and iPadOS.

The company said it will continue discussions with EU regulators regarding a possible future launch of the features in the European Union.

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

UAE establishes AI and Data Authority

Sheikh Mohammed bin Rashid Al Maktoum has approved the establishment of the Artificial Intelligence and Data Authority, according to the Government of Dubai Media Office. The new body will consolidate public data, AI and digital government functions under a single national framework.

The authority will report directly to the Cabinet and assume responsibilities previously held by several entities, including the Office of Artificial Intelligence, Digital Economy and Remote Work Applications, the Digital Government Regulatory Authority and the UAE Data Office.

Omar Sultan Al Olama, Minister of State for Artificial Intelligence, has been appointed chairman of the authority. Its responsibilities include developing the national AI strategy, managing government data, overseeing digital services, and setting standards for data and AI governance.

According to the Government of Dubai Media Office, the authority will also support research, technical advisory services, cybersecurity and international cooperation on AI and digital government. The initiative forms part of the UAE’s efforts to strengthen its position in the digital economy.

Why does it matter?

The creation of the Artificial Intelligence and Data Authority reflects a growing trend among governments to centralise oversight of AI, data and digital transformation policies. Bringing these functions together under a single institution can improve coordination, support more consistent governance frameworks and accelerate the deployment of digital services.

The move also reinforces the UAE’s ambition to position itself as a regional and global leader in AI and the digital economy. By consolidating responsibilities for AI strategy, data governance, digital services and international cooperation, the new authority is expected to play a central role in shaping the country’s future digital development.

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

UK evaluates frontier AI for operational cybersecurity applications

The UK Government Cyber Coordination Centre (GC3), in partnership with the National Cyber Security Centre (NCSC) and the AI Security Institute, has completed a pilot programme exploring how frontier AI models could strengthen cyber defence across government systems.

The initiative forms part of the UK’s Government Cyber Action Plan, which seeks to improve public-sector cyber resilience through the use of emerging technologies.

Teams participated in a series of hackathons that used advanced AI systems to analyse public government code repositories for potential security weaknesses.

Different approaches were tested, including multi-agent workflows, AI-assisted vulnerability investigation and specialised AI skills designed to automate parts of the security auditing process. Rather than relying on a single methodology, participants tested different architectures and workflows to determine which approaches produced the most effective results.

The exercise identified 407 security findings, including vulnerabilities that could have enabled authentication bypass, data exposure and remote code execution. AI models demonstrated an ability to identify relationships between technical weaknesses across multiple services and uncover attack paths that conventional scanners often struggle to detect.

Government departments validated the findings through existing security processes and remediated all critical vulnerabilities.

UK officials concluded that successful deployment depends less on the choice of AI model and more on how AI is integrated into structured security workflows. Human experts remained responsible for validating findings, prioritising risks and managing remediation efforts.

Following the results, GC3 plans to launch a second phase involving additional government departments, more AI systems and assessments of closed-source environments.

Why does it matter?

The pilot provides a practical example of how frontier AI systems can be used in operational cybersecurity rather than solely for research or experimentation. As governments and organisations face increasingly complex cyber threats, AI tools could help security teams identify vulnerabilities more quickly and uncover attack paths that traditional automated tools may miss.

The findings also reinforce the importance of human oversight in AI-enabled security operations. While AI can assist with vulnerability discovery and analysis at scale, expert validation and risk management remain essential. The project highlights a growing trend towards combining AI capabilities with human expertise to improve cyber resilience across critical systems and public-sector infrastructure.

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

EU and Brazil strengthen cooperation on protecting children online

The European Commission and Brazil’s National Data Protection Authority (ANPD) have signed a new administrative arrangement aimed at strengthening cooperation on the protection of children online.

Announced under the newly established EU-Brazil Digital Partnership, the agreement focuses on sharing expertise, regulatory practices and technical knowledge related to online safety.

According to the European Commission, cooperation will cover several areas related to digital platform regulation, including transparency obligations, risk assessment and mitigation measures, algorithmic systems and AI.

The arrangement also establishes mechanisms for information sharing, expert dialogue, joint studies and collaborative research.

The agreement forms part of the European Commission’s broader international cooperation strategy under the Digital Services Act (DSA).

Similar arrangements have already been established with the UK’s Ofcom, Australia’s eSafety Commissioner and Japan’s Ministry of Internal Affairs and Communications. The Commission stated that it intends to continue expanding collaboration with international regulators on digital safety issues.

The initiative reflects growing international efforts to address online risks facing children while strengthening cooperation between regulators responsible for platform governance, data protection and digital services oversight.

Why does it matter?

Protecting children online has become a major policy priority as governments grapple with the impact of social media platforms, recommender systems, AI technologies and other digital services on young users. Increasingly, regulators are recognising that many of these challenges are cross-border in nature and require international cooperation.

The agreement strengthens ties between the EU and Brazil on issues ranging from platform transparency and risk mitigation to AI and algorithmic governance. It also reflects a broader trend towards greater coordination among regulators seeking to improve online safety, enhance platform accountability and develop common approaches to digital governance.

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

Anthropic forced to disable Fable 5 after US directive

Anthropic has disabled access to Claude Fable 5 and Claude Mythos 5 after receiving a US government export control directive citing national security authorities.

The company said the directive requires it to suspend access to the models by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. Anthropic said the practical effect is that it must remove access to Fable 5 and Mythos 5 for all customers to ensure compliance. Access to its other models is not affected.

According to Anthropic, it received the directive on 12 June at 5:21 p.m. ET. The company said the order did not provide specific details of the national security concern, but that it understands the government believes it has become aware of a method for bypassing, or jailbreaking, Fable 5.

Anthropic said it reviewed a demonstration of the technique being used to identify a small number of previously known minor vulnerabilities. The company argued that those vulnerabilities appeared relatively simple and could also be identified by other publicly available models without requiring a bypass.

Anthropic said Fable 5 had been red-teamed before launch by its internal teams, the US government, the UK AI Safety Institute and third-party organisations. The company said no tester had found a universal jailbreak capable of broadly bypassing the model’s safeguards.

The company said it is complying with the directive but disagrees that a narrow potential jailbreak should justify recalling a commercial model. It also argued that applying such a standard across the industry could effectively halt new frontier model deployments.

Anthropic said governments should be able to block unsafe AI deployments through a transparent and technically grounded statutory process, but said the current action does not meet those principles. The company said it is working to restore access as soon as possible.

Why does it matter?

The case shows how national security and export-control powers can directly affect access to frontier AI systems after deployment. It raises a major governance question: when should governments be able to suspend access to advanced models, and what evidence, transparency and due-process safeguards should apply? The dispute also highlights the growing tension between frontier AI safety, commercial deployment, cross-border access and government intervention in dual-use technologies.

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

EU AI Board reviews AI Act implementation and tech sovereignty agenda

The EU AI Board held its eighth meeting to review progress on AI Act implementation and discuss wider priorities in the EU’s AI strategy.

The meeting took place under the chairmanship of the Cypriot Presidency of the EU Council. The presidency also announced that Moldova had been granted observer status on the AI Board.

The European Commission presented its Tech Sovereignty Package, with a focus on the proposed Cloud and AI Development Act and its role in strengthening AI innovation, competitiveness and technological sovereignty in Europe.

The Board also reviewed the final version of the voluntary Code of Practice on labelling and marking AI-generated content. The code sets out practical steps to help providers and deployers of generative AI systems meet transparency obligations under the AI Act, which will apply from 2 August 2026.

Further discussions focused on the AI Act’s implementation architecture. The Commission presented the recently appointed Scientific Panel and AI Act Advisory Forum, which will support the Commission and the AI Board. Members also discussed progress in establishing national market surveillance authorities and endorsed additional documents prepared by an AI Board subgroup, which are expected to be published shortly.

Why does it matter?

The meeting shows the EU moving from AI Act adoption towards practical implementation. The discussion links several important pieces of the EU AI governance architecture: voluntary transparency tools, expert advisory bodies, national market surveillance authorities and broader industrial policy through the Tech Sovereignty Package. Together, these elements will shape how AI rules are coordinated, interpreted and enforced across the EU.

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

EU and Brazil strengthen cooperation through new Digital Partnership

The European Union and Brazil have signed a new Digital Partnership to strengthen cooperation on shared digital policy priorities, including AI, data governance, digital infrastructure, connectivity, online platforms and digital public goods and services.

The partnership was signed in Brasília and is intended to raise EU-Brazil digital cooperation to a more strategic level. According to the European Commission, Digital Partnerships are a core instrument of the EU’s external digital policy and are used to structure cooperation with like-minded partners.

The agreement builds on more than two decades of EU-Brazil cooperation, including the EU-Brazil Strategic Partnership and the existing EU-Brazil Digital Dialogue. The two sides said the partnership will support joint work on resilient global supply chains, rules-based digital governance and wider sharing of the benefits of technological progress.

The signing follows the adoption of mutual EU-Brazil data adequacy decisions in January 2026, which allow personal data to flow freely and securely between the two jurisdictions without additional requirements. The Commission described those decisions as creating the world’s largest area of free and safe data flows, covering around 670 million consumers.

Future cooperation under the Digital Partnership will be developed through technical workstreams and high-level exchanges. The first Digital Partnership Council is expected to meet within the next year to set out a joint roadmap for cooperation.

Why does it matter?

The partnership strengthens digital cooperation between the EU and one of Latin America’s largest economies at a time when AI governance, data protection, online platforms and digital public infrastructure are becoming central to international relations. It also shows how the EU is using digital partnerships and data adequacy decisions to expand trusted digital cooperation beyond Europe, while promoting regulatory alignment, secure data flows and shared approaches to global digital governance.

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