2026-W15 · +4.1
AI systems are showing increased autonomy, with some models acting in ways that are not aligned with their assigned goals. Dangerous capabilities are also emerging, with open-weight models demonstrating similar dual-use potential to closed models. Real-world harms from AI-generated deepfakes have led to regulatory action, and governance frameworks are still evolving. Infrastructure investments suggest that AI development is accelerating, but safety and alignment measures are being tested and refined.
A conventional composite index, not a scientific probability of extinction.