Skip to content

Responses & Interventions

This section covers what can be done about AI risk—from technical research to policy interventions to building societal resilience. Understanding the landscape of possible responses helps individuals, organizations, and policymakers decide where to focus.


Direct work on making AI systems safer:

CategoryDescription
Technical ApproachesAlignment research, interpretability, evaluations, AI control
Research AgendasSpecific research programs and their theories of change

Policy, regulation, and coordination:

CategoryDescription
LegislationEnacted and proposed laws (EU AI Act, US EO, etc.)
Compute GovernanceHardware-based mechanisms like export controls, compute monitoring
InternationalSummits, treaties, and international coordination
Industry Self-RegulationRSPs, voluntary commitments, industry standards

Organizations and bodies working on AI governance:

CategoryDescription
AI Safety InstitutesGovernment bodies focused on AI safety
Standards BodiesOrganizations developing AI standards

Technologies and mechanisms for collective intelligence:

CategoryDescription
Epistemic ToolsPrediction markets, forecasting, verification systems
Coordination TechnologiesTools for large-scale cooperation

Growing capacity and preparing for challenges:

CategoryDescription
Field BuildingGrowing the AI safety community
ResilienceBuilding societal capacity to handle AI disruption

When evaluating interventions, consider:

  • How much risk reduction if successful?
  • Does it address a critical bottleneck?
  • Is meaningful progress possible?
  • What’s the track record?
  • How much work is already happening?
  • What’s the marginal value of more resources?

Your beliefs about AI risk affect which interventions look most promising:

If you believe…Prioritize…
Short timelines (<5 years)Governance, existing systems, immediate impact
Alignment is very hardTechnical research, fundamental breakthroughs
Misuse is the main riskAccess controls, monitoring, defensive capabilities
Racing dynamics dominateInternational coordination, industry agreements
Institutions work wellPolicy advocacy, standards development

Different backgrounds enable different contributions:

BackgroundPotential Contributions
TechnicalAlignment research, interpretability, evaluations
PolicyGovernance research, legislative work, standards
OperationsSupporting research organizations
CommunicationsPublic engagement, translating research
FundingGrantmaking, donor advising