Longterm Wiki

Risks

Directory of AI-related risks tracked in the knowledge base, including accident risks, misuse risks, structural risks, and epistemic risks.

Total Risks
65
Accident
44
Misuse
8
Structural
6
Epistemic
7
With Severity
62
With Likelihood
62
Showing 65 of 65 risks
Risk▲CategorySeverityLikelihoodTime Horizon
AI Authoritarian ToolswikiMisuseHighHigh~2025
AI Capability SandbaggingwikiAccidentHighMedium2025–2030
AI Development Racing DynamicswikiStructuralHighHigh~2025
AI DisinformationwikiMisuseHighVery High~2025
AI Distributional ShiftwikiAccidentMediumVery High~2025
AI Flash DynamicswikiAccidentHighMedium High~2025
AI Knowledge MonopolywikiAccidentCriticalMedium2030–2050
AI Mass SurveillancewikiMisuseHighVery High~2025
AI Model SteganographywikiAccident———
AI Preference ManipulationwikiAccidentHighMedium2025–2035
AI ProliferationwikiAccidentHighHigh~2025
AI Surveillance and US Democratic ErosionwikiAccidentHighHigh2025–2028
AI Trust Cascade FailurewikiEpistemicCriticalMedium2025–2040
AI Value Lock-inwikiStructuralCatastrophicMedium2030–2045
AI Winner-Take-All DynamicswikiStructuralHighHigh~2025
AI-Accelerated Reality FragmentationwikiAccidentHighMedium2025–2035
AI-Driven Concentration of PowerwikiStructuralHighMedium High2025–2040
AI-Driven Economic DisruptionwikiStructuralMedium HighHigh~2030
AI-Driven Institutional Decision CapturewikiAccidentHighMedium2025–2040
AI-Driven Legal Evidence CrisiswikiAccidentHighMedium2025–2035
AI-Driven Trust DeclinewikiEpistemicMedium HighHigh~2025
AI-Enabled Authoritarian TakeoverwikiAccidentCatastrophicMedium2025–2050
AI-Enabled Biological RiskswikiAccidentCatastrophicMedium2025–2032
AI-Enabled CyberattackswikiAccidentHighHigh~2025
AI-Enabled Historical RevisionismwikiAccidentHighMedium2025–2040
AI-Enabled Untraceable MisusewikiAccidentHighHigh~2025
AI-Induced Cyber PsychosiswikiAccidentMedium HighMedium2025–2030
AI-Induced EnfeeblementwikiStructuralMedium HighMedium~2030
AI-Induced Expertise AtrophywikiAccidentHighMedium2025–2050
AI-Induced IrreversibilitywikiAccidentCriticalMedium~2030
AI-Powered Consensus ManufacturingwikiEpistemicHighMedium2025–2030
AI-Powered DeanonymizationwikiAccidentHighHigh~2025
AI-Powered FraudwikiMisuseHighVery High~2025
AI-Powered Investigation RiskswikiAccidentMediumHigh~2025
Authentication CollapsewikiEpistemicCriticalMedium2025–2030
Automation Bias (AI Systems)wikiEpistemicMediumVery High~2025
Autonomous ReplicationwikiAccidentHighMedium2025–2030
Autonomous WeaponswikiMisuseHighHigh~2025
Bioweapons RiskwikiMisuseCatastrophicMedium2025–2030
Compute ConcentrationwikiAccidentHighHigh2025–2035
Concentrated Compute as a Cybersecurity RiskwikiAccidentHighMedium2025–2035
Corrigibility FailurewikiAccidentCatastrophicHigh~2035
Cyberweapons RiskwikiMisuseHighHigh~2025
Deceptive AlignmentwikiAccidentCatastrophicMedium~2035
DeepfakeswikiMisuseMedium HighVery High~2025
Emergent CapabilitieswikiAccidentHighMedium~2025
Epistemic CollapsewikiEpistemicHighMedium High~2030
Epistemic Learned HelplessnesswikiAccidentHighMedium2030–2050
Epistemic SycophancywikiEpistemicMedium HighMedium2025–2030
Erosion of Human AgencywikiAccidentMedium HighHigh~2030
Financial Stability Risks from AI Capital ExpenditurewikiAccidentMediumMedium2026–2032
Goal MisgeneralizationwikiAccidentHighHigh2025–2030
Instrumental ConvergencewikiAccidentHighHigh~2035
Key Near-Term AI RiskswikiAccident———
Mesa-OptimizationwikiAccidentCatastrophicMedium~2035
Multipolar Trap (AI Development)wikiAccidentHighMedium High~2030
Power-Seeking AIwikiAccidentCatastrophicMedium~2035
Reward HackingwikiAccidentHighVery High~2025
Rogue AI ScenarioswikiAccidentCatastrophicMedium2026–2040
SchemingwikiAccidentCatastrophicMedium~2035
Scientific Knowledge CorruptionwikiAccidentHighMedium2024–2035
Sharp Left TurnwikiAccidentCatastrophicMedium~2035
Sleeper Agents: Training Deceptive LLMswikiAccident———
SycophancywikiAccidentMediumVery High~2025
Treacherous TurnwikiAccidentCatastrophicMedium~2035