Artificial intelligence Archives | Page 7 of 677

Australian Finance Department releases internal AI guidance

Australia’s Department of Finance has publicly released internal guidance on generative AI under the Freedom of Information Act 1982, outlining how staff and contractors should use AI tools in their work.

The guidance, dated March and April 2026, applies to tools including Microsoft 365 Copilot Chat, Microsoft 365 Copilot full licences, and public generative AI services such as ChatGPT, Claude, and Gemini. It says AI tools can improve productivity and service delivery, but also carry risks that must be understood and managed.

Staff intending to use AI tools must complete the APS Academy’s AI in Government Fundamentals course. Staff are also encouraged to build prompting skills in a secure environment through GovAI’s Interactive Learning Environment and discuss approved AI use cases with managers.

The guidance says staff must use generative AI safely, responsibly, and ethically, in line with departmental policies and APS values. The guidance states that AI should support, rather than replace, human judgement and that final decisions must always be made by people, not AI systems.

The document also sets limits on the information that can be entered into AI tools. Public generative AI tools may be used with non-sensitive official and unofficial information, but staff must not enter personal, sensitive, classified or protected information. Copilot Chat and Copilot full licence in web mode are also restricted from use with personal, sensitive, classified, or protected information.

Finance’s enterprise-grade Copilot full licence operating in work mode permits broader use within the department’s ICT environment, including sensitive and protected information where an AI use case has been formally registered. Staff seeking to use personal information with Copilot full licence must complete a Privacy Impact Threshold Assessment and consult the Privacy Team.

The guidance also requires staff to register some AI use cases through Finance’s AI Use Case Register. Registration is required for certain uses involving personal, sensitive, classified, or protected information, while paid AI tools or systems require consultation or approval before procurement.

Staff are required to disclose AI use when AI-generated content significantly influences decision-making, could reasonably be mistaken for human-generated content, has not been reviewed by a subject-matter expert, or where legal or ethical obligations require disclosure.

The department says AI use is overseen by an AI Governance Committee responsible for promoting AI strategy, supporting safe implementation, advising on ethical, legal and social responsibilities, and ensuring compliance with government legislation, regulations and standards.

The guidance says Finance governs AI in line with the Digital Transformation Agency’s Policy for the responsible use of AI in government, Australia’s AI Ethics Principles, the Pilot Australian Government AI assurance framework, and the Protective Security Policy Framework. It says the department has limited its use of AI to low-risk use cases.

Why does it matter?

The guidance provides a practical example of how governments are translating high-level AI principles into operational rules for everyday use. Rather than focusing solely on ethics frameworks, it addresses concrete issues such as training requirements, approved tools, data handling, disclosure obligations and governance processes.

The document also highlights a broader challenge facing public administrations worldwide. As generative AI becomes part of routine government work, agencies must balance productivity gains with privacy, security, transparency and accountability requirements. Australia’s approach illustrates how governments are seeking to enable AI adoption while maintaining human oversight and limiting risks associated with sensitive information.

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

EDPS warns Shadow AI creates hidden data protection risks

The European Data Protection Supervisor (EDPS) has warned that Shadow AI can create hidden data protection and breach risks when employees use unauthorised AI tools without organisational approval. The warning was published in a blog post by EDPS Wojciech Wiewiórowski on 15 June 2026.

The EDPS said Shadow AI can include tools such as generative AI chatbots, coding assistants and automated note-taking applications. While employees may use them as shortcuts to improve productivity, unauthorised AI tools can bypass data protection and security safeguards.

According to the EDPS, data entered into unapproved AI tools can fall into a regulatory and compliance blind spot. Unauthorised tools may lack formal agreements governing the legal basis for processing, data retention periods and safeguards for international data transfers.

The EDPS also warned that Shadow AI can create a transparency gap, making it difficult for organisations to determine where information is stored, how it is processed or whether it is used to train AI models. Such tools can also introduce security vulnerabilities, including automated meeting recorders joining meetings without oversight from IT security teams.

The blog post argues that organisations should address these risks proactively rather than attempting to ignore or prohibit them outright. Instead, they should adopt proactive AI governance policies that define authorised AI use, establish data classification rules and set approval processes for new technologies.

The EDPS said policies should be backed by technical controls and monitoring, including blocking unapproved AI domains, enforcing data loss prevention rules and restricting the installation of unauthorised AI software. The EDPS also recommended that organisations provide approved AI platforms that are secure, compliant and capable of meeting employees’ operational needs.

The EDPS said reducing Shadow AI risks requires cooperation between data protection officers, IT departments, security teams and business functions. The aim, it said, is to protect data subject rights and institutional information while enabling responsible AI adoption.

Why does it matter?

Shadow AI turns everyday workplace AI use into a data protection and cybersecurity issue. Employees may use unauthorised tools to save time, but organisations can lose visibility over personal data, legal compliance, retention, international transfers and model training.

The warning also shows that responsible AI adoption depends on more than staff guidance. Organisations need approved AI tools, technical controls, monitoring and cooperation between data protection, IT, security and business teams to reduce breach risks without blocking useful innovation.

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

Armenian finance minister highlights AI’s economic potential and risks

Armenia’s Finance Minister Vahe Hovhannisyan said AI could support economic growth while also creating new economic and labour-market challenges. He made the comments during a parliamentary discussion on the performance of the 2025 state budget.

Hovhannisyan said the impact of AI is being widely debated internationally and that governments around the world are actively exploring its economic implications. He was responding to questions about AI’s potential effect on GDP growth and the expansion of the tax base.

The minister cited international estimates suggesting that AI adoption could add approximately 0.8 to 1 percentage point to economic growth. He said AI has the potential to generate new forms of employment while supporting productivity and economic growth.

At the same time, Hovhannisyan warned that AI could disrupt existing jobs and create adjustment challenges for labour markets. The remarks were made during discussions on Armenia‘s 2025 budget performance, as the government’s 2026 budget projects economic growth of 5,4%.

Why does it matter?

The comments reflect a broader global debate about AI’s economic impact. Policymakers increasingly view AI as a potential driver of productivity, innovation and economic growth, while also recognising the possibility of labour-market disruption and changing workforce demands.

For emerging economies such as Armenia, the challenge is not only adopting AI technologies but also ensuring that workers and businesses can benefit from them. The long-term impact of AI on growth, employment and public finances will depend on investment, skills development and the ability to adapt to technological change.

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

Canada seeks stronger privacy rights through new digital governance law

The Canadian government has introduced the Protecting Privacy and Consumer Data Act, a major legislative proposal designed to modernise the country’s private-sector privacy framework and strengthen protections in an increasingly AI-driven digital environment.

According to the government, Canada’s existing privacy legislation was developed more than 25 years ago and no longer reflects technological realities such as AI, automated decision-making systems, deepfakes and the large-scale collection of children’s data.

The proposed law seeks to address those challenges by establishing stronger rights for individuals and clearer obligations for organisations.

The legislation would recognise privacy as a fundamental right, strengthen protections for children’s data, require meaningful consent for the collection and use of personal information, and introduce greater transparency around automated decision-making.

Canadians would also gain the right to request the deletion of their personal information and benefit from enhanced safeguards when their data is transferred outside Canada.

The proposed framework would be overseen by a newly established Digital Safety and Data Protection Commission of Canada.

The regulator would have authority to issue binding orders and impose significant penalties on organisations that fail to comply with privacy requirements. The government describes the legislation as a key component of its recently launched national AI strategy, aimed at strengthening trust in digital services, data-driven innovation and emerging technologies.

Why does it matter?

The proposed legislation represents one of Canada’s most significant privacy reforms in decades and reflects a broader international trend of updating data protection frameworks for the AI era. As AI systems, automated decision-making tools and digital platforms become more deeply embedded in everyday life, governments are seeking stronger safeguards for personal data, transparency and accountability.

The bill also signals a growing convergence between privacy policy and AI governance. By introducing stronger protections for children’s data, new rights for individuals and greater oversight of automated systems, Canada is positioning privacy as a key foundation for public trust in digital services and emerging technologies.

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

OECD says governments need stronger delivery capacity for digital transformation

The OECD says governments have made progress in building the foundations of digital government, but must now focus on turning those foundations into measurable benefits for people and businesses.

In its Digital Government Outlook 2026, the OECD says governments are operating under pressure from rapid technological change, fiscal constraints, rising public expectations and the growing adoption of AI. The report argues that digital technologies and data are now essential to public-sector performance, resilience, and trust.

The Outlook draws on the 2025 OECD Digital Government Index and the Open, Useful and Re-usable Data Index. It covers 36 OECD members and eight accession candidate countries, including Argentina, Brazil, Bulgaria, Croatia, Indonesia, Peru, Thailand, and Romania.

The report finds that OECD countries have strengthened key digital foundations, including shared infrastructure, interoperable systems, digital identity, cloud services and open data frameworks. The average Digital Government Index score rose from 0.61 in 2023 to 0.70 in 2025, while the OURdata Index increased from 0.48 to 0.53.

However, the OECD says progress remains uneven. Countries tend to perform better in setting strategic direction and policy frameworks than in implementation and monitoring. The report says governments often have strategies and enabling mechanisms in place but struggle to embed them in day-to-day operations, workflows and accountability systems.

AI adoption is one of the main areas where this gap is visible. The OECD says AI is already used in at least one area of government in almost every OECD country, and most countries have strategies, oversight bodies, and training programmes. Yet only 28% of OECD countries systematically assess the financial and non-financial impacts of AI use in government.

The report also points to gaps in digital skills and investment evaluation. Only six OECD countries have a dedicated strategy for developing digital skills among civil servants, while just one in four systematically evaluates whether completed digital projects delivered their intended results.

The OECD says the next phase of digital government should focus on wider adoption of interoperable systems, stronger data governance, more strategic investment and skills development, trustworthy AI at scale, and more joined-up, user-centred public services. The OECD argues that governments must move beyond fragmented digital initiatives and embed digital technologies, data and AI into everyday public-sector operations.

Why does it matter?

The report suggests that the challenge facing digital government is no longer primarily technological. Many governments have already established digital identities, cloud infrastructure, interoperable systems and data frameworks. The next challenge is ensuring these foundations translate into better public services, greater efficiency and stronger public trust.

The findings also highlight a growing implementation gap in areas such as AI. While governments are increasingly adopting AI tools and digital technologies, many lack the skills, evaluation frameworks and governance mechanisms needed to measure outcomes and scale successful initiatives. As a result, the effectiveness of future digital government reforms may depend less on technology adoption and more on institutional capacity and execution.

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

European Commission opens applications for RAISE AI research advisory board

The European Commission has opened applications for the RAISE High-Level Academic Advisory Board, inviting leading researchers in AI and AI-enabled science to help shape Europe’s future AI research agenda.

The advisory board will support the implementation of the EU’s AI in Science Strategy and provide independent scientific guidance on the development of RAISE (Resource for AI Science in Europe).

RAISE was launched in 2025 under Horizon Europe to strengthen European leadership in both fundamental AI research and the application of AI across scientific disciplines.

The Commission is seeking academics with expertise in AI research or experience applying AI in fields such as medicine, climate science and advanced materials. Board members will provide strategic recommendations on research priorities, governance structures, benchmarks and framework conditions needed to accelerate AI-enabled scientific discovery.

Through RAISE, the EU aims to bring together leading researchers, computational resources, data and funding within a coordinated ecosystem that supports scientific excellence and strengthens Europe’s position in global AI research and innovation.

Why does it matter?

The initiative reflects growing recognition that AI is becoming a foundational tool for scientific discovery across disciplines ranging from healthcare and climate research to materials science and physics. Governments are increasingly investing in AI research infrastructure to ensure that researchers have access to the computing power, data and expertise needed to remain globally competitive.

The advisory board also highlights Europe’s ambition to play a larger role in shaping the future of AI-enabled science. By coordinating talent, infrastructure and funding through initiatives such as RAISE, the EU aims to strengthen both its scientific capacity and its position in the global race for AI innovation.

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

Humanists UK urges government to adopt human-centred AI principles

Humanists UK has urged the UK government to place human dignity, democratic oversight and human flourishing at the centre of AI governance.

The call followed a House of Lords debate on the impact of AI on human relationships and society, during which peers discussed the ethical, social and regulatory challenges raised by rapidly advancing AI systems.

Humanists UK pointed out to the government the Luxembourg Declaration on Artificial Intelligence and Human Values, adopted by Humanists International in 2025. The declaration argues that AI should support human judgement, the common good, democratic governance, transparency, autonomy and protection from harm.

Lord Michael Cashman, a patron of Humanists UK and member of the All-Party Parliamentary Humanist Group, urged the government not to ‘reinvent the wheel’ and said the declaration already sets out principles relevant to AI governance.

Liberal Democrat peer Lord Clement-Jones said the debate showed a convergence of values across different traditions, including the need for democratic oversight, transparency and safeguards to ensure AI serves human beings rather than replacing them.

Responding for the government, Digital Economy Minister Baroness Lloyd of Effra said AI is already changing the economy, public services and human relationships. She said the government’s responsibility is to ensure that the transformation strengthens rather than diminishes the fabric of society.

Humanists UK said it has written to Baroness Lloyd and shared a copy of the Luxembourg Declaration.

Why does it matter?

The story reflects the growing role of civil society, religious groups and ethical movements in AI governance debates. While it does not signal a new UK policy, it shows how discussions on AI safety are broadening beyond technical risk to include human dignity, democratic accountability, transparency, autonomy and the public interest. Such value-based frameworks may influence how governments frame future AI regulation, assurance and safeguards.

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

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!