This policy report examines how ethical, legal, and technical interoperability can reduce fragmentation in AI safety governance, drawing on comparative country studies from China, South Korea, Singapore, and the United Kingdom.
An independent, evidence-led synthesis of global research on risks from general-purpose and frontier AI systems, focusing on capability growth, misuse, systemic risk, and the state of AI risk management practices worldwide.
A concise, practitioner-oriented guide explaining how organizations can extend an existing ISO 27001 Information Security Management System to cover AI-specific risks and governance requirements under ISO/IEC 42001:2023.
“AI readiness extends beyond technology. Strategy, organisation, data, and technology together determine whether AI-enabled services can be deployed responsibly and at scale.”
Cybersecurity in 2026 is accelerating amid growing threats, geopolitical fragmentation and a widening technological divide. Artificial intelligence is reshaping both defence and offence, testing governance, skills and global cooperation.
A growing consensus has emerged worldwide that the future of AI must be shaped thoughtfully and responsibly, ensuring that innovation enhances human well-being while safeguarding fundamental rights.
A flagship brief mapping how global AI laws, standards, and voluntary frameworks converge and diverge, identifying risk management, transparency, and security-by-design as emerging universal anchors of AI governance.
A practical governance framework designed to help organizations harness generative AI while managing emerging risks across strategy, data, operations, ethics, and accountability.
A structured checklist by the German Federal Office for Information Security outlining how to understand, model, and mitigate evasion attacks on LLM systems, including prompt injections, jailbreaks, and agent-based attack paths.
This guidance provides EU institutions with a structured framework to identify and mitigate data protection risks arising from the development, procurement, and deployment of AI systems processing personal data.
MAS proposes sector-wide Guidelines on AI Risk Management to set supervisory expectations for financial institutions, covering AI oversight, lifecycle controls, risk materiality, governance structures, and capabilities, including Generative AI and AI agents.
A constitutional approach to AI alignment that defines high-level principles guiding model behavior, prioritizing safety, helpfulness, and respect for human values through transparent, rule-based self-governance.
This white paper proposes a joint framework that bridges AI ethics with practical governance and accountability mechanisms, drawing on lessons from Africa and the Asia-Pacific to move from high-level principles to real-world execution.
A practical technical standard issued by the Australian Government to support the safe, responsible, and transparent design, development, deployment, and operation of AI systems across public sector use cases.
This report analyses how high-risk AI systems are assessed in practice under the EU AI Act, focusing on fundamental rights risks, current gaps, and the need for practical guidance for providers and deployers.
The artificial intelligence (AI) impact assessment tool helps Australian Government teams identify, assess and manage AI use case impacts and risks against Australia’s AI Ethics Principles.
A comprehensive analysis of how global AI governance is emerging, fragmenting, and evolving, using regime complexity, sociotechnical change, and governance disruption as core analytical lenses.
AI agents are adaptive, interactive systems that must be continuously evaluated, secured, governed, and improved due to the nondeterministic nature of underlying LLMs.
AI capabilities are improving rapidly across all tested domains. Performance in some areas is doubling every eight months, and expert baselines are being surpassed rapidly.
A comprehensive framework by the Cloud Security Alliance proposing four integrated pillars—Model Cards, Data Sheets, Risk Cards, and Scenario Planning—to proactively identify, assess, and mitigate risks across the AI/ML lifecycle.
AI literacy is the ability to understand, evaluate, and confidently use AI technologies, recognising both their capabilities and limitations in personal, professional, and societal contexts.
This white paper presents a comprehensive framework for AI incident response, adapting proven practices from complex systems domains to address non-deterministic, context-dependent, and interconnected AI failures.
AI is increasingly used by foresight practitioners to accelerate horizon scanning, trend analysis and scenario development, but it raises significant challenges around trust, bias, transparency and governance.
An analytical framework for understanding AI governance proposals by surfacing their shared and divergent assumptions, helping policymakers act under uncertainty while preserving future decision-making flexibility.
Curated Library of AI Governance Resources
