Week of 29 June 2026
FLARE-AI is an open-source system enabling standardised, multi-recipient AI flaw and incident reporting via a single submission.
Key points
- Developed with 49 experts across 32 organisations including Anthropic, Google, MITRE, CERT, and major incident databases.
- Australia has no equivalent coordinated AI flaw disclosure infrastructure; this framework could inform future APS approaches.
MIT AI Risk Repository tested eight LLMs against human expert reviewers for classifying AI incidents across five taxonomies.
Key points
- Opus 4.6, with targeted prompt refinement, matched human-baseline agreement on all five taxonomies including EU AI Act risk levels.
- Findings are methodologically useful for APS teams considering LLM-assisted classification or incident monitoring pipelines.
Week of 20 April 2026
MIT AIRI's new Navigator tool unifies AI risk, incident, governance, and mitigation datasets under a shared taxonomy.
Key points
- Policymakers can explore how governance documents map to specific risk domains against real-world incident records.
- Governance data skews toward US sources, limiting direct applicability to Australian regulatory contexts.
Week of 6 April 2026
MIT AI Risk Repository maps over 1,000 governance documents, revealing gaps in socioeconomic risk and early lifecycle coverage.
Key points
- Findings show governance documents concentrate on model safety, public administration, and downstream lifecycle stages - potentially relevant for APS gap analysis.
- Dataset is heavily US-federal in origin, limiting direct applicability to Australian governance landscape without supplementary analysis.
Week of 23 February 2026
EPIC's 2023 framework identifies nine harm categories from generative AI, from physical injury to dignitary harm.
Key points
- The MIT AI Risk Repository has catalogued this as its 31st AI risk framework - a growing reference library for governance practitioners.
- Framework is US-origin and advocacy-driven; useful for taxonomy comparison but not directly calibrated to Australian regulatory context.
Week of 16 February 2026
A 2023 academic paper proposes a taxonomy of 7 major LLM trustworthiness categories covering 29 subcategories.
Key points
- The MIT AI Risk Repository spotlights this as one of 30 risk frameworks it has catalogued - useful for APS risk inventory work.
- The paper itself is two years old; the blog post adds no new analysis beyond the repository spotlight.
MIT AI Risk Repository spotlights the AI TRiSM framework covering trust, risk, and security management across AI lifecycles.
Key points
- Framework organises AI risks under bias, privacy, deepfakes, societal manipulation, autonomous weapons, and malicious use.
- This is a literature synthesis blog post - the underlying 2024 academic paper carries more analytical depth.
Week of 9 February 2026
MIT AI Risk Repository spotlights a 2023 safety taxonomy for Chinese LLMs covering 8 harm scenarios and 6 adversarial attack types.
Key points
- The taxonomy claims scalability beyond Chinese-language models, making it potentially relevant to broader LLM safety evaluation work.
- This is a blog summary of a 2023 academic paper - useful reference material, not new guidance or policy.
SafetyBench is a bilingual benchmark assessing LLM safety across 7 risk categories using 11,435 multiple-choice questions.
Key points
- The MIT AI Risk Repository spotlights this as one of 28 frameworks cataloguing AI risks - useful for comparative evaluation work.
- A 2023 academic paper; this blog post adds no new findings beyond summarising the original arXiv publication.
Week of 2 February 2026
MIT AI Risk Repository spotlights a Google DeepMind-led paper on ethical risks of advanced AI assistants.
Key points
- Framework covers value alignment, human-assistant interaction risks, and societal-scale impacts across three structured areas.
- Identifies an 'evaluation gap' where current approaches focus on model-level considerations rather than broader sociotechnical effects.
A 2023 paper proposes embedding model evaluation for dangerous capabilities and alignment into AI governance processes.
Key points
- Nine dangerous capability categories are identified, including cyber-offense, deception, self-proliferation, and situational awareness.
- MIT AI Risk Repository surfaces this as one of 25 risk frameworks - useful reference material for agencies building AI risk taxonomies.
Singapore's AI Verify Foundation developed an 11-principle testing framework covering transparency, safety, fairness, and accountability.
Key points
- The framework aligns with ASEAN, EU, OECD, and US AI governance frameworks, giving it cross-jurisdictional reference value.
- This item is a MIT AI Risk Repository blog spotlight - the substantive content originates from a 2023 Singapore Government document.
Week of 22 December 2025
MLCommons AI Safety Benchmark v0.5 defines 13 hazard categories for evaluating chat-based AI system safety.
Key points
- Practical testing tools including ModelBench are openly available, making this usable for agency-level AI evaluation.
- V0.5 has been superseded by V1.0 (AILuminate, Feb 2025); this spotlight is retrospective context, not a new release.
MIT AI Risk Repository spotlights a 2023 paper categorising catastrophic AI risks into four proximate causes.
Key points
- The four categories — malicious use, AI race, organisational risks, and rogue AI — each include mitigations.
- This is a secondary blog summary of a 2023 paper; primary value is as a reference for risk taxonomy work.
Week of 15 December 2025
A 2022 academic framework organises AI/ML risks into data-level and model-level categories with root causes and outcomes.
Key points
- The framework targets high-stakes decision settings like healthcare and transport - domains relevant to APS service delivery.
- This is a 2022 paper spotlighted by MIT's AI Risk Repository blog; it is not new primary research or Australian guidance.
Week of 1 December 2025
MIT AI Risk Repository Version 4 now includes over 1,700 coded risks drawn from 74 published frameworks.
Key points
- Nine newly added frameworks span government reports, peer-reviewed papers, and industry sources, including a UK DSIT frontier AI paper.
- A structured, living reference for AI risk taxonomy - useful for APS governance and risk assessment work.
Week of 13 October 2025
MIT AI Risk Repository used LLMs to classify 950+ AI governance documents across risk, mitigation, and sector taxonomies.
Key points
- Governance failure, security vulnerabilities, and transparency were the most-covered risk domains; AI welfare and multi-agent risks were least covered.
- US-heavy dataset limits global generalisability; Australian documents are unlikely to be well-represented in current outputs.
Week of 25 August 2025
MIT AI Risk Repository publishes a transparent, publicly accessible deck of 13 AI risk mitigation frameworks.
Key points
- The resource consolidates academic, industry, and policy sources into a draft AI Risk Mitigation Taxonomy for governance use.
- Useful reference for APS teams building or auditing internal AI risk frameworks, though not Australia-specific.
Week of 28 July 2025
MIT AI Risk Repository extracted 831 mitigations from 13 frameworks into a searchable database with a four-category taxonomy.
Key points
- The taxonomy covers Governance & Oversight, Technical & Security, Operational Process, and Transparency & Accountability controls - directly mapping to APS AI governance concerns.
- Operational Process Controls and Testing & Auditing were the most frequently cited mitigations; Model Alignment was rarely mentioned despite its importance.
Week of 14 July 2025
MIT AI Risk Repository's incident tracker has been updated to include all AIID incidents through 23 June 2025 (up to ID #1116).
Key points
- New features include national security impact assessment across five categories, harm severity rescaling, and Fishbone causal diagrams.
- Useful as a reference dataset for APS agencies developing AI risk registers or incident classification frameworks.
A 2022 academic framework proposes six AI risk categories specifically designed for public sector governance contexts.
Key points
- The taxonomy links technological, ethical, legal, social, economic, and informational risks to concrete governance guidelines.
- MIT AI Risk Repository blog spotlight - the underlying paper is three years old and the signal is retrospective rather than new.
MIT AI Risk Repository spotlights a 2020 academic framework organising AI governance challenges for public administration into three categories.
Key points
- The framework's five-layer governance structure and four-stage regulatory process offer a reference model for agency AI risk management.
- The underlying paper is five years old; APS practitioners likely have more current frameworks already in use.
Week of 12 May 2025
MIT AI Risk Repository spotlights a 2022 Google DeepMind taxonomy of LLM risks across six domains and 20 subdomains.
Key points
- The taxonomy covers discrimination, information hazards, misinformation, malicious use, HCI harms, and socioeconomic harms - directly relevant to APS AI risk assessment work.
- The underlying paper is from 2022; the MIT blog post is a summary spotlight, not new research.
Week of 21 April 2025
MIT's AI Risk Repository updated to 1,612 unique risk entries across 65 frameworks, now including multi-agent risks.
Key points
- The repository provides causal and domain taxonomies designed to support policy, auditing, and governance processes.
- A credible reference resource for APS agencies developing AI risk frameworks or audit criteria - freely accessible.
MIT AI Risk Repository spotlights the 2022 MLSR framework, categorising ML risks into first-order and second-order types.
Key points
- The framework offers a structured taxonomy integrating impact assessments, incident reports, and ML literature - useful for risk assessment design.
- This is a 2022 academic paper being surfaced via a blog digest; it is reference material rather than new guidance.