Longterm Wiki

Dangerous Capability Evaluations

Systematic testing of AI models for dangerous capabilities including bioweapons assistance, cyberattack potential, autonomous self-replication, and persuasion/manipulation abilities to inform deployment decisions and safety policies. Now standard practice with 95%+ frontier model coverage.

Related

Related Pages

Top Related Pages

Safety Research

Anthropic Core Views

Risks

Emergent CapabilitiesAI Proliferation

Analysis

AI Uplift Assessment ModelBioweapons Attack Chain Model

Approaches

AI EvaluationAlignment EvaluationsAI Safety Cases

Organizations

Apollo ResearchSurvival and Flourishing FundUK AI Safety InstituteGoogle DeepMind

Policy

AI Safety Institutes (AISIs)

Concepts

Alignment Evaluation Overview

Other

Stuart RussellBeth BarnesAI EvaluationsRed Teaming

Key Debates

AI Misuse Risk CruxesAI Safety Solution Cruxes

Tags

dangerous-capabilitiesbioweaponscybersecurityself-replicationdeployment-decisionsresponsible-scaling