The UK AI Safety Institute, renamed to "AI Security Institute" on February 14, 2025 by Science Secretary Peter Kyle at Munich Security Conference. Focus shifted from broad existential/alignment safety to near-term national security risks: cybersecurity, CBRN, fraud, child exploitation. New criminal misuse team (joint with Home Office) and partnership with Defence Science and Technology Laboratory (Dstl). Signed MoU with Anthropic to explore AI in public services. Established 2023 at Bletchley Park AI Safety Summit.
AI EvaluationApproachAI EvaluationComprehensive overview of AI evaluation methods spanning dangerous capability assessment, safety properties, and deception detection, with categorized frameworks from industry (Anthropic Constituti...Quality: 72/100AI Governance Coordination TechnologiesApproachAI Governance Coordination TechnologiesComprehensive analysis of coordination mechanisms for AI safety showing racing dynamics could compress safety timelines by 2-5 years, with $500M+ government investment in AI Safety Institutes achie...Quality: 91/100
Analysis
Intervention Timing WindowsAnalysisIntervention Timing WindowsFramework for prioritizing AI safety interventions by temporal urgency rather than impact alone, identifying four critical closing windows (2024-2028): compute governance (70% closure by 2027), int...Quality: 72/100AI Safety Research Allocation ModelAnalysisAI Safety Research Allocation ModelAnalysis finds AI safety research suffers 30-50% efficiency losses from industry dominance (60-70% of ~$700M annually), with critical areas like multi-agent dynamics and corrigibility receiving 3-5...Quality: 65/100
Policy
Bletchley DeclarationPolicyBletchley DeclarationThe Bletchley Declaration represents a significant diplomatic achievement in establishing international consensus on AI safety cooperation among 28 countries including the US and China, though its ...Quality: 60/100AI Safety Institutes (AISIs)PolicyAI Safety Institutes (AISIs)Analysis of government AI Safety Institutes finding they've achieved rapid institutional growth (UK: 0→100+ staff in 18 months) and secured pre-deployment access to frontier models, but face critic...Quality: 69/100
Organizations
AnthropicOrganizationAnthropicComprehensive reference page on Anthropic covering financials ($380B valuation, $14B ARR at Series G growing to $19B by March 2026), safety research (Constitutional AI, mechanistic interpretability...Quality: 74/100
Concepts
International Coordination MechanismsConceptInternational Coordination MechanismsComprehensive analysis of international AI coordination mechanisms shows growing but limited progress: 11-country AI Safety Institute network with ~$200M budget expanding to include India; Council ...Quality: 91/100Compute ThresholdsConceptCompute ThresholdsComprehensive analysis of compute thresholds (EU: 10^25 FLOP, US: 10^26 FLOP) as regulatory triggers for AI governance, documenting that algorithmic efficiency improvements of ~2x every 8-17 months...Quality: 91/100Autonomous CodingCapabilityAutonomous CodingAI coding capabilities reached 70-76% on curated benchmarks (23-44% on complex tasks) as of 2025, with 46% of code now AI-written and 55.8% faster development cycles. Key risks include 45% vulnerab...Quality: 63/100Situational AwarenessCapabilitySituational AwarenessComprehensive analysis of situational awareness in AI systems, documenting that Claude 3 Opus fakes alignment 12% baseline (78% post-RL), 5 of 6 frontier models demonstrate scheming capabilities, a...Quality: 67/100
Other
AI EvaluationsResearch AreaAI EvaluationsEvaluations and red-teaming reduce detectable dangerous capabilities by 30-50x when combined with training interventions (o3 covert actions: 13% → 0.4%), but face fundamental limitations against so...Quality: 72/100Ian HogarthPersonIan HogarthAI investor and writer; former Chair of the UK AI Safety Institute (2023-2025).Red TeamingResearch AreaRed TeamingRed teaming is a systematic adversarial evaluation methodology for identifying AI vulnerabilities and dangerous capabilities before deployment, with effectiveness rates varying from 10-80% dependin...Quality: 65/100Oliver SourbutPersonOliver SourbutResearcher and AI specialist at the Future of Life Foundation. Former UK AI Safety Institute researcher. Engaged with OECD, UK FCDO, and DSIT on AI governance. Researched agent oversight at Oxford....
Historical
AI Safety Summit (Bletchley Park)HistoricalAI Safety Summit (Bletchley Park)International summits convening governments and AI labs to address AI safety0
Risks
Bioweapons RiskRiskBioweapons RiskComprehensive synthesis of AI-bioweapons evidence through early 2026, including the FRI expert survey finding 5x risk increase from AI capabilities (0.3% → 1.5% annual epidemic probability), Anthro...Quality: 91/100AI Knowledge MonopolyRiskAI Knowledge MonopolyAnalyzes the risk that 2-3 AI systems could dominate humanity's knowledge access by 2040, projecting 80%+ market concentration with correlated errors and epistemic lock-in. Provides comprehensive m...Quality: 50/100
Key Debates
AI Accident Risk CruxesCruxAI Accident Risk CruxesComprehensive survey of AI safety researcher disagreements on accident risks, quantifying probability ranges for mesa-optimization (15-55%), deceptive alignment (15-50%), and P(doom) (5-35% median ...Quality: 67/100AI Governance and PolicyCruxAI Governance and PolicyComprehensive analysis of AI governance mechanisms estimating 30-50% probability of meaningful regulation by 2027 and 5-25% x-risk reduction potential through coordinated international approaches. ...Quality: 66/100