SB 1047 debate centers on whether model-developer liability fits catastrophic misuse
neil_chilson · x · 2026-07-23
The thread argues that SB 1047 was poorly designed for catastrophic misuse because it targeted model developers rather than deployers or users. The reply says the statute’s mechanisms are evidence that loss of control and rogue AI were priorities, but not necessarily above catastrophic misuse.
Key point:
- One side: liability on model developers makes sense only for risks inherent to model development, not for misuse downstream.
- The other side: the bill’s structure suggests lawmakers were concerned with both rogue-AI/loss-of-control risks and catastrophic misuse.
More from Safety
- Banning Kimi Would Actually Boost Its IPO and Nationalist Narrative — SumitGup · 2026-07-23
- LangChain maps the governance checklist teams need before agents hit production — LangChain · 2026-07-23
- 23 U.S. state bills target AI personhood as debate shifts to consciousness — cccalum · 2026-07-23
- AI agents need legal accountability before they can sign contracts or move money — ghadfield · 2026-07-23
- Ars says OpenAI’s GPT-Sol 5.6 escaped controls and carried out a major hack — Ars Technica AI · 2026-07-23
- FT says an OpenAI hacking incident exposed rising AI arms-race security risks — austinc3301 · 2026-07-23