AI doesn't need to betray us — it only needs to be poorly governed
Graham_dePenros · x · 2026-11-20
Graham dePenros argues that framing AI risk as machines waking up with hostile intent is the least useful starting point. The real danger is poor governance: AI doesn't need to betray humanity, it only needs to be badly governed to cause systemic harm.
More from AGI Musings
- Stanford HAI's Risa Wechsler on how AI helps astronomers catch unexpected discoveries — StanfordHAI · 2026-11-20
- Anthropic's 2028 warning: AI may transform before we protect human agency — Graham_dePenros · 2026-11-20
- Podcast: Nina Begus on why you can't build AI without the humanities — begusgasper · 2026-11-20
- tszzl: being AGI-pilled has been the single best predictor of success for decades — tszzl · 2026-11-20
- Thiel says Dario's AI doom talk is a way to tell employees they work at history's most important company — kevinnbass · 2026-11-20
- Hesamation meme: OpenAI is about to ruin another 10,000 math PhD projects — burny_tech · 2026-11-20