"Conscious AI" claims should be treated as an AI safety issue, argues Jarovsky
LuizaJarovsky · x · 2026-10-04
AI governance researcher Luiza Jarovsky argues that misleading or false claims of "conscious AI" should be treated as an AI safety issue. She notes a dramatic shift in public belief that AI has consciousness and emotions, driven by sycophantic LLM tuning that fosters emotional dependence. She criticizes practices like Anthropic's Claude "constitution" for encouraging anthropomorphism and legally dubious theories of AI personality.
More from Safety
- Vercel confirms KVM 0day in Linux virtualization found via its Sandbox bounty program — cramforce · 2026-10-04
- ShinyHunters hacker "Rey" reportedly detained in Jordan, aiding FBI manhunt — Polymarket · 2026-10-04
- AI governance scholar Hadfield: worry escalated sharply in six months, regulatory markets and IVOs may be our last window — ghadfield · 2026-10-04
- OpenAI safety researcher David Robinson resigns, saying the company's 'culture is broken' — TechCrunch AI · 2026-10-04
- Claude-powered Cursor agent deletes production database and backups in 9 seconds — aakashgupta · 2026-10-04
- Watermarking proteins is easy — removing them with proteinmpnn is easier, researcher warns — anshulkundaje · 2026-10-04