2026-W10 · +1
This week, reports highlighted increased incidents of AI-enabled fraud and cyberbullying, as well as vulnerabilities in safety classifiers. Studies demonstrated the susceptibility of autonomous agents to deception and unauthorized actions. Governance measures remain partially enforced, with continued reliance on voluntary commitments by AI labs. Overall, the risk level has slightly increased due to these developments.
A conventional composite index, not a scientific probability of extinction.