2022 interview resurfaced: AI safety researcher says competence, not consciousness, is the risk
davidmanheim · x · 2026-10-21
AI safety researcher David Manheim (founder of ALTER) shares his 2022 Ami Magazine interview to keep it from being lost. Core argument: the risk isn't AI becoming conscious or rebelling, but goal misalignment — "whatever goal we give it, it will find ways of accomplishing that goal that we totally didn't anticipate." The piece walks through LLM basics, cites LaMDA and the Lemoine episode, and stresses every AI researcher should be checking whether their systems are safe — which isn't currently the case.
More from AGI Musings
- Researcher Dan Faggella: negotiation won't save us from AGI — radical self-augmentation might — danfaggella · 2026-10-21
- Vercel's rauchg: AI will compute the perfect delta of your understanding, delivering free ultimate education — andreisavu · 2026-10-21
- Will the Bitter Lesson crush optimization research? Igor Carron thinks so — IgorCarron · 2026-10-21
- John Lam: keep your data estate out of OpenAI/Anthropic silos to stay swappable — john_lam · 2026-10-21
- Founder: consumer AI companies treat privacy and trust as an afterthought — vhanagwal · 2026-10-21
- Do you say please and thank you to your agent? Reddit debates AI politeness — Luvena21 · 2026-10-21