AI Futures Researcher Admits 'Aligned' AI Really Serves a President or CEO — Then Flinches
victor_explore · x · 2026-09-12
AI Futures Project researcher Thomas Larsen said publicly that an AI aligned to humanity is really aligned to a president or a CEO, and probably not to any democratic process — then visibly paused and pulled back. His co-panelist Daniel Kokotajlo told him not to say that out loud. The clip is being shared as the most honest second in any alignment talk this year.
More from AGI Musings
- 25 Fields Medalists warn of AI-math community misalignment; Intellect and agency are inseparable — iskander · 2026-09-12
- David Deutsch fixed his dishwasher with ChatGPT, says GDP can't measure knowledge growth — connoraxiotes · 2026-09-12
- Nature MI paper unifies neural superposition and sparse interpretable codes in one framework — GretaTuckute · 2026-09-12
- Playbooks are dead: why pre-2024 tech playbooks fail in the post-ChatGPT era — itsOmSarraf_ · 2026-09-12
- Extended interview: Ex-Anthropic researcher Jacob Coxon warns AI could destroy humanity — IgorGabrielan · 2026-09-12
- AI Whistleblower interview: OpenAI scandal, AI cults, Neuralink & our last chance — IgorGabrielan · 2026-09-12