AI Control Risks Prompt Legislative Safeguards and Research
AI 안전 | Fri Jul 24 2026 00:00:00 GMT+0000 (Coordinated Universal Time) | 3 sources
US Congress prepares an AI kill switch bill, researchers publish an omnicidal risk taxonomy, and DARPA demonstrates AI-piloted F-16 flights.
Analysis
[US Congress] prepared to introduce the AI Kill Switch Act [1]
- Bipartisan introduction by Representatives Ted Lieu and Nathaniel Moran
- Grants DHS authority to order AI system shutdowns
- Targets loss-of-control scenarios
- Violations carry fines up to $20M per day
[arXiv (Andrew Critch et al.)] published a taxonomy of AI omnicidal risks [2]
- Presents a classification framework for human extinction scenarios
- Frames events as avoidable possibilities rather than inevitabilities
- Aims to support preventive measures through public discourse
[DARPA & U.S. Air Force] demonstrated autonomous flight of an AI-piloted F-16 [3]
- Modified F-16 as autonomous flight testbed under the VENOM program
- Human-on-the-loop switch toggles between human and AI control
- Multi-agent AI live flight testing planned under the AIR program