• Anthropic released its first annual “Public Benefit Report” which sets out its progress and its gaps against five charter goals, from making AI safe to sharing its benefits widely.

• HM Treasury’s new Data Centre Covenant sets five tests for data centres, covering energy costs, community benefit, environmental standards, local jobs and national security, with faster grid access and planning support for strategic sites that meet them.

• Drawing on new data from 40 countries, the OECD’s “Government Unstuck: The Next Frontier of Government Transformation” identifies practices in digital technology, AI, citizen participation, regulation and procurement that can help governments become simpler, more agile and more trusted.

• A new Eurobarometer survey, Workplace-related drivers and obstacles for the adoption of Artificial Intelligence and Algorithmic Management by SMEs and Mid-Caps, looks at how European small and mid-sized businesses are using AI to manage their workforce, and at what helps or holds back wider adoption.

• In “From Digital Scale to Data Power: A Data Operating Model for India’s States” the Tony Blair Institute argues that India’s states need clear data ownership, shared standards and independent quality checks, so the data from India’s large digital systems can be trusted and used for better services and safe AI.

• CivicPlus’s “The State of Resident Trust in Local Government” report finds that trust in US local government has risen since 2022, and residents who use their council’s website more than once a month are nearly five times more likely to report complete trust.

• In “Democratic Governance of Frontier AI, Part II: A Blueprint for Empowering Cyber Defenders,” OpenAI calls on governments to tighten safety and security rules for advanced AI, modernise cyber defences, and strengthen public institutions that test AI and respond to incidents.

• The OECD policy brief “Securing AI Compute: Government Policy Approaches” reviews how governments are approaching access to AI computing power, building on its 2023 blueprint for national compute capacity.

• The World Economic Forum and Bain & Company report “Unlocking Value: AI for Resilient Growth in MENA” draws on five AI prototypes built with partners in the region to show how AI can support productivity, sustainable finance and climate resilience in the Middle East and North Africa.

• In “Determining Public Interest in the Secondary Use of Health Data” the OECD sets out five groups of criteria for deciding when reusing health data serves the public interest: population health, research and innovation, ethics and risk, health equity, and transparency.

• The World Economic Forum white paper “Unlocking Azerbaijan’s TradeTech Potential” sets out 10 recommendations to help Azerbaijan use digital tools in trade, including legal recognition of digital trade documents and better links between government and business platforms.

• The OECD policy brief “Enhancing User Control and Privacy Protections in the Context of AI” looks at how to give people more control over their personal data when it is collected through chatbot conversations and web scraping, and recommends closer work between regulators and industry.

• The GovAI-led technical report “Managing Biological Risks from General-Purpose AI Models under the EU AI Act“ makes nine recommendations to help AI providers and the EU AI Office assess biological risks from advanced AI models more fully and consistently.

• The Brookings briefing “How Are Countries Evaluating AI Models?” reviews how governments and independent groups test AI models, finding broad agreement that independent testing is needed but open questions over what to measure, who should do it, and how results should be used.

• The WEF white paper “Powering AI at Scale: Optimizing Hyperscale Data Centre Infrastructure in MENA” offers a framework for expanding large data centres in the Middle East and North Africa without overloading power and water systems, through coordinated action on energy, cooling and regulation.

• Based on more than 200 initiatives, the OECD’s “Data Spaces between Co-ordination Challenges, Financial Viability and Public Policy” finds that few data spaces have become financially sustainable despite public investment, and it offers eight policy considerations to help them get started.

Articles:

• “Are Language Models Rational? The Case of Coherence Norms and Belief Revision” argues that rules of rational belief apply to some language models but not to others, which matters for predicting their behaviour and for AI safety.
• “Citizens’ Perceptions of Accountability in AI-Based Public Sector Decision Making” uses a nationally representative US survey to explore whom citizens think should be held accountable for AI-based government decisions, and what information about those decisions should be shared with the public.
• “The Paradox of Experience: Public Procurement and Delivery Timelines across Digital Solution Types” studies more than 40,000 digital projects in Italian municipalities. It finds that past experience in buying digital solutions speeds up some projects but slows down others, such as cloud migration and citizen-facing services.
• “AI Infrastructure, Governance Preparedness, and Digital State Systems in Africa” finds that more integrated AI infrastructure in Africa tends to go hand in hand with weaker governance conditions, and that data protection laws alone give only a partial picture of a country’s readiness.
• “Large Language Models for Machine Translation Quality Annotation: Humans and Models Are Both Challenged” finds that neither AI models nor human reviewers judge translation quality reliably in most settings, and suggests that people and AI working together may give better results.
• “Artificial Intelligence for Public Use” argues that AI could help ease labour shortages and raise productivity in Sweden’s public sector, but that the benefits would vary widely across different parts of the sector.
• “Optimizing for Happiness in EU Technology Policy to Repay Public Subsidies for AI” argues that EU policymakers should require AI companies receiving public support to show measurable benefits for people’s wellbeing, and hold them to account for delivering them.

Blogs:

• Writing for the IAPP, Kris Klein examines how regulators in Canada can make responsible AI practices more visible and measurable, strengthening transparency, accountability and public trust in AI governance.
• A GovInsider column explores why some governments advance digital transformation faster than others, highlighting the importance of institutional flexibility, coordinated decision-making and political leadership.
• In a World Economic Forum opinion piece, Etay Maor argues that organisations should apply zero-trust principles and clear accountability to AI systems, treating them as digital insiders rather than ordinary software.
• At the UN Security Council, African leaders called for a direct role in setting global AI safety standards, warning that the continent should not become a testing ground for AI models built elsewhere.

• An IAPP round-up covers key developments in India’s digital policy landscape, including efforts to strengthen data protection, AI governance and online safety while balancing innovation with regulatory oversight.

• Rest of World reports how the rise of autonomous AI agents could deepen digital inequalities, particularly in developing countries facing infrastructure, accessibility and regulatory challenges.
• A World Economic Forum article argues that technology is moving faster than the rules that govern it, and that closing gaps in governance, digital access, cybersecurity and energy use needs cooperation across sectors.

• The UK AI Security Institute explores emerging best practices for frontier AI development, highlighting the importance of robust risk assessments, continuous monitoring and rigorous safety evaluations.

• EU Technological Sovereignty Has an Unanswered Question: What Do We Want It For? argues that the EU’s technological sovereignty package focuses on chips, data centres and reducing dependencies but has no funding behind it and leaves citizens little say in what kind of digital future Europe should build.

• GovInsider examines how technical jargon can exclude citizens from digital public infrastructure design, emphasising the need for accessible communication and more inclusive public participation.

• Funding: UKRI Enabling AI adoption across science and engineering, UK Research and Innovation (UKRI)

• AI4Gov-X Cascade Funding Call, AI4GovAccelerate

• Call for evidence: Vision for an AI-enabled clean energy system, Department for Energy Security and Net Zero (DESNZ)

• Open call for guest investigators in the life sciences, Anthropic

• Call for evidence: Agentic AI, Information Commissioner’s Office (ICO)

• Call for Proposals: New approaches for Human/AI collaboration for the workforce of the future, European Commission (Horizon Europe)