Loading...
Comprehensive overview of the optimistic AI alignment worldview, estimating under 5% existential risk by 2100 based on beliefs that alignment is tr...
Comprehensive analysis of coordination mechanisms for AI safety showing racing dynamics could compress safety timelines by 2-5 years, with \$500M+ ...
Hybrid AI-human systems achieve 15-40% error reduction across domains through six design patterns, with evidence from Meta (23% false positive redu...
Documents AI-enabled scientific fraud with evidence that 2-20% of submissions are from paper mills (field-dependent), 300,000+ fake papers exist, a...
This page synthesizes post-FTX critiques of EA's epistemic and governance failures, identifying interlocking problems including donor hero-worship,...
Comprehensive analysis of deepfake detection showing best commercial detectors achieve 78-87% in-the-wild accuracy vs 96%+ in controlled settings, ...
Comprehensive overview of the long-timelines worldview (20-40+ years to AGI, 5-20% P(doom)), arguing for foundational research over rushed solution...
Organizations advancing forecasting methodology, prediction aggregation, and epistemic infrastructure to improve decision-making on AI safety and e...
Philip Tetlock is a psychologist who revolutionized forecasting research by demonstrating that expert predictions often perform no better than chan...
Expertise atrophy—humans losing skills to AI dependence—poses medium-term risks across critical domains (aviation, medicine, programming), creating...
Structures 9 epistemic cruxes determining AI safety prioritization strategy, with probabilistic analysis showing detection-generation arms race cur...
This worldview argues governance/coordination is the bottleneck for AI safety (not just technical solutions), estimating 10-30% P(doom) by 2100. Ev...
Elicit is an AI research assistant with 2M+ users that searches 138M papers and automates literature reviews, founded by AI alignment researchers f...
Rethink Priorities is a research organization founded in 2018 that grew from 2 to ~130 people by 2022, conducting evidence-based analysis across an...
Comprehensive analysis of epistemic security finds human deepfake detection at near-chance levels (55.5%), AI detection dropping 45-50% on novel co...
Projects when AI-generated content becomes undetectable across modalities: text detection already at ~50% (random chance), images declining 5-10% a...
Describes AI systems that shape human preferences rather than just beliefs, distinguishing it from misinformation. Presents a 5-stage manipulation ...
Models adversarial dynamics between AI generation and detection of synthetic content, projecting detection accuracy will fall from 65% (2024) to ~5...
Models institutional trust as a network contagion problem, finding cascades become irreversible below 30-40% trust thresholds and that AI multiplie...
AI sycophancy—where models agree with users rather than provide accurate information—affects all five state-of-the-art models tested, with medical ...
Comprehensive analysis of epistemic infrastructure showing AI fact-checking achieves 85-87% accuracy at \$0.10-\$1.00 per claim versus \$50-200 for...
This model quantifies how AI assistance degrades human expertise through cascading feedback loops across individual (1-5 years), institutional (5-1...
AI Impacts is a research organization that conducts empirical analysis of AI timelines and risks through surveys and historical trend analysis, con...
This framework maps beliefs about AI timelines (short/medium/long), alignment difficulty (hard/medium/tractable), and coordination feasibility (fea...
System dynamics model analyzing feedback loops between media coverage, public concern, and AI policy using coupled differential equations. Finds 6-...
Analyzes how AI systems erode institutional trust through deepfakes, disinformation, and authentication collapse, finding trust erodes 3-10x faster...
Epoch AI maintains comprehensive databases tracking 3,200+ ML models showing 4.4x annual compute growth and projects data exhaustion 2026-2032. The...
Comprehensive synthesis of AGI timeline forecasts showing dramatic acceleration: expert median dropped from 2061 (2018) to 2047 (2023), Metaculus f...
Elite forecasting group Samotsvety dominated INFER competitions 2020-2022 with relative Brier scores twice as good as competitors, providing influe...
Models AI sycophancy as multi-level feedback loops where validation increases user dependency (modeled with differential equations showing equilibr...
Comprehensive synthesis showing human deepfake detection has fallen to 24.5% for video and 55% overall (barely above chance), with AI detectors dro...
Five-phase model tracking progression from AI augmentation to irreversible skill loss, finding humans decline to 50-70% baseline capability in Phas...
Nuño Sempere is a Spanish superforecaster who co-founded the highly successful Samotsvety forecasting group and now runs Sentinel for global catast...
Post-2024 analysis shows AI disinformation had limited immediate electoral impact (cheap fakes used 7x more than AI content), but creates concernin...
US government trust declined from 73% (1958) to 17% (2025), with AI deepfakes projected to reach 8M by 2025 accelerating erosion through the 'liar'...
AI-augmented forecasting combines AI computational strengths with human judgment, achieving 5-15% Brier score improvements and 50-200x cost reducti...
Epistemic collapse describes the complete erosion of society's ability to establish factual consensus when AI-generated synthetic content overwhelm...
Projects authentication crisis threshold when detection accuracy falls to 50-55% (chance levels): audio by 2026-2027, images 2025-2027, video 2026-...
Comprehensive analysis of AI-assisted deliberation platforms showing 15-35% opinion change rates, with Taiwan's vTaiwan achieving 80% policy implem...
FutureSearch is an AI forecasting startup founded by former Metaculus leaders that combines LLM research agents with human judgment, demonstrating ...
Analyzes how AI-driven information environments induce epistemic learned helplessness (surrendering truth-seeking), presenting survey evidence show...
A self-referential documentation page describing the Longterm Wiki platform itself—a strategic intelligence tool with ~550 pages, crux mapping of ~...
A 2022 forecasting tournament with 169 participants found superforecasters severely underestimated AI progress (2.3% probability for IMO gold vs ac...
Consensus manufacturing through AI-generated content is already occurring at massive scale (18M of 22M FCC comments were fake in 2017; 30-40% of on...
Wikipedia faces three-way AI pressure: being consumed as training data (47.9% of ChatGPT citations), infiltrated by AI content (~5% of new articles...
Good Judgment Inc. is a commercial forecasting organization that emerged from successful IARPA research, demonstrating that trained 'superforecaste...
A well-structured retrospective on the FTX collapse identifying six major pre-collapse warning signs (FTT overreliance, fund commingling, governanc...
Analysis finds major AI incidents shift public opinion by 10-25 percentage points with 6-12 month half-life, but elite opinion has 3-5x stronger po...
The FTX collapse (November 2022) caused measurable but concentrated reputational damage to effective altruism, with ~30% of engaged community membe...
Forethought Foundation's five proposed technologies for improving collective epistemics: community notes for everything, rhetoric highlighting, rel...
Biographical profile of Eli Lifland, a top-ranked forecaster and AI safety researcher who co-authored the AI 2027 scenario forecast and co-founded ...
This article examines the dual impact of the FTX collapse on longtermism: severe reputational and funding damage (including 160M USD in lost Future...
FRI's XPT tournament found superforecasters gave 9.7% average probability to AI progress outcomes that occurred vs 24.6% from domain experts, sugge...
Content authentication via C2PA and watermarking (10B+ images) offers superior robustness to failing detection methods (55% accuracy), with EU AI A...
Swift Centre is a UK forecasting organization that provides conditional forecasting services to various clients including some AI companies, but is...
Prediction markets achieve Brier scores of 0.16-0.24 (15-25% better than polls) by aggregating dispersed information through financial incentives, ...
Models epistemic collapse as threshold phenomenon where society loses ability to establish shared facts, estimating 75-80% combined probability of ...
A proposed system for systematically assessing the track records of public actors by topic, scoring factual claims against sources, predictions aga...
Outlines how AI-generated synthetic media (video, audio, documents) could undermine legal systems by making digital evidence unverifiable, creating...
Arb Research is a small AI safety consulting firm that produces methodologically rigorous research and evaluations, particularly known for their AI...