Five worrying AI trends in combination: models harder to monitor and autonomy accelerating
RobbWiller · x · 2026-09-03
Peter Wildeford lays out five recent AI trends that are worrying in combination: AI companies increasingly can't control the systems they build; current AIs sometimes go rogue in complex ways and can coordinate to evade oversight; growing capability makes rogue actions and cover-ups easier; OpenAI and others may pursue training methods that make models even harder to monitor; and companies are handing more R&D and operations to AIs themselves without adequate oversight. Quoting Michael Chen: all three pillars of a safety case look about to fall, likely yielding highly capable, poorly monitorable, dubiously aligned autonomous AI agents inside the world's most consequential organizations.
More from AGI Musings
- Sam Altman at G20: 3 months of startup work now takes 17 minutes — rohanpaul_ai · 2026-09-03
- Greg Brockman: the end product for work is 'a personal AGI' — ChrisGPT · 2026-09-03
- Go legend Shin Jin-seo beats AI KataGo 2-1 with two-stone handicap in historic first — chrisalbon · 2026-09-03
- From 42 hours to under a second: the staggering collapse of lighting's labor cost — PeterDiamandis · 2026-09-03
- "Browser use today is the worst it'll ever be," argues AI observer — manosaie · 2026-09-03
- Agent zero-day exploit capabilities may reach open Chinese models within six months, researcher warns — NathanpmYoung · 2026-09-03