This page is still being polished. If you have thoughts, please share them via the feedback form.
Data on this page is preliminary and may change. Please do not share or cite these figures publicly.
Response plans, escalation procedures, and incident management for operational emergencies.
Also in Operations & Security
Stage: Escalation; Stakeholder: Third Party Researchers; Additional information: Government stakeholders should establish disclosure channels and whistleblower safeguards for employees of AI developers. Employees across key sectors -- including AI developers, compute providers and third-party evaluators -- should have designated channels to report safety concerns, as well as protections against potential retaliation or suppression, including the use of non-disclosure agreements (NDAs), non-disparagement clauses or other measures to suppress disclosure of safety risks (Lovely 2024). In addition, governments should create secure disclosure channels that are independent from corporate protocols and make these available to employees at developer firms.
Reasoning
Escalation procedure communicates identified risks to external researchers for awareness and response coordination.
Monitor critical capability levels
2.2.2 Testing & EvaluationIdentify early warning signs and emergent capabilities
2.2.1 Risk AssessmentEstablish standardised benchmarks and reporting
3.2.1 Benchmarks & EvaluationImplement compute monitoring and anomaly detection
1.2.3 Monitoring & DetectionEnhance hardware and supply chain oversight
2.3.3 Monitoring & LoggingLead efforts to establish shared criteria for AI LOC
3.2.2 Technical StandardsStrengthening Emergency Preparedness and Response for AI Loss of Control Incidents
Somani, Elika; Friedman, Anjay; Wu, Henry; Lu, Marianne; Byrd, Christopher; van Soest, Henri; Zakaria, Sana (2025)
As artificial intelligence (AI) systems become increasingly embedded in essential infrastructure and services, the risks associated with unintended failures rise. Developing comprehensive emergency response protocols could help mitigate these significant risks. This report focuses on understanding and addressing AI loss of control (LOC) scenarios where human oversight fails to adequately constrain an autonomous, general-purpose AI.
Operate and Monitor
Running, maintaining, and monitoring the AI system post-deployment
Other
Actor type not captured by the standard categories
Govern
Policies, processes, and accountability structures for AI risk management
