WHAT THE AI SAFETY & CYBER SECURITY TRACKER COVERS

Return to the AI — Safety & Cyber Security Tracker

AI SAFETY & CONTROL

• Loss-of-control warnings from researchers, executives and public officials
• Unexpected autonomous behavior and attempts to evade oversight
• Alignment, deception, containment and self-improvement evidence
• Guardrail failures, removals and circumvention
• Safety-team departures, insider warnings and institutional disputes
• Changes to laboratory safety policies and deployment thresholds
• Government regulation, evaluations and international coordination

AI CYBER SECURITY

• AI-enabled intrusion, reconnaissance and vulnerability discovery
• Autonomous or semi-autonomous cyber operations
• Agent swarms and machine-speed attack campaigns
• Defensive-system failures and successful containment
• AI model theft, manipulation and supply-chain compromise
• Attacks on critical infrastructure, financial systems and military networks
• Cybersecurity standards, reporting requirements and regulatory responses