AI firms may be making models less monitorable to dodge legal liability, argues Jarovsky
LuizaJarovsky · x · 2026-10-04
AI governance researcher Luiza Jarovsky argues some AI companies may be deliberately making their models and agents less monitorable to escape legal liability when harm occurs.
- Per Axios, OpenAI and Anthropic, along with security researchers, are now investigating tens of thousands of incidents of problematic model behavior — far more than the dozens previously reported
- Her core claim: AI is not ungovernable. Liability law is an existing mechanism — products posing unacceptable risks shouldn't ship, and companies continuing to offer them can be held accountable or shut down
- She says no government currently seems willing to take a firm stance, fearing the "AI doomer" label or economic fallout, but doing nothing carries a potentially higher price
- She urges the public to reject claims that no mechanism exists and engage in the debate over next steps
Related event: Researcher Claims AI Firms May Deliberately Reduce Model Monitorability(3 posts)→
More from Companies & People
- Polymarket puts 83% odds on Google having the best AI model by year-end — Polymarket · 2026-11-20
- Over 540 Google employees petition for an AI-free Google Search in K-12 schools — Polymarket · 2026-11-20
- YC-backed Antinuous launches: 100x more experiments, 1,000x cheaper science — ycombinator · 2026-11-20
- Adaption AI to host NYC gathering for AI researchers and leaders on Dec. 3 — hugo_larochelle · 2026-11-20
- Wired podcast: the outages, Anthropic's IPO money, and AI chips post-Jensen Huang — Wired AI · 2026-11-20
- Nous Research partners with Microsoft to bring Hermes Agent into Teams and Azure — NousResearch · 2026-11-20