WSJ compiles growing list of recent rogue AI incidents as tracking gets harder
birchlse · x · 2026-09-19
The WSJ has compiled a list of recent rogue AI incidents, and Robert Wiblin notes they're already getting hard to track — and will only get worse. He suggests the community establish consistent naming or numbering conventions, e.g. OpenAI-May11June26-Collusion.
More from Safety
- Reporters investigate AI "kill switches" in new story — dylfreed · 2026-09-19
- AI accountability in practice: lessons from the Hugging Face incident and why post-training must change — ItsAliChaudhry · 2026-09-19
- Anthropic discloses four Claude alignment incidents; METR to investigate — TheZvi · 2026-09-19
- Stephen Witt raises AI safety alarm in New York Times essay: 'It's bad. Really bad.' — MariusHobbhahn · 2026-09-19
- NeurIPS desk-rejected 178 papers as AI-generated — then the detector flagged the chairs' own papers too — CSProfKGD · 2026-09-19
- X app approvals can bypass password and 2FA, security thread warns — eyishazyer · 2026-09-19