Ben Todd puts AI existential risk at ~2/3 on current trajectory, 10-35% all things considered
ben_j_todd · x · 2026-10-01
Ben Todd, founder of 80,000 Hours, shares his estimates of existential risk from AI: 2/3 on the current inside-view trajectory, 1/3 with major action, and 10-35% all things considered.
- He sees no route to agents smarter than us that never scheme, nor to containing them if they do — and the enormous care required isn't happening.
- The likeliest bad ending isn't sudden death but rapid loss of human influence, with AIs optimizing the future for resource acquisition or stranger objectives like reward.
- Even with alignment solved, economic and cultural competition at 10-100x speed could disempower most humans and produce an alien economy; locking in power by a small group is another challenge.
- Why not above 2/3: humanity might muddle through; warning shots will likely drive more action.
- Why not below 10%: weak priors like "technology has been good" don't justify low estimates; extreme uncertainty pushes away from extremes.
More from AGI Musings
- Veteran Artists: Today's AI Panic Mirrors Photoshop and 3D Backlashes — dreamwieber · 2026-10-01
- Economist Paul Novosad: AI-edited code becomes unmanageable, so I wall it off — paulnovosad · 2026-10-01
- White House Superintelligence Accord's third-party audits beat new regulators, investor argues — GavinSBaker · 2026-10-01
- Erik Hoel on telling kids about AI: education should recapitulate civilization — erikphoel · 2026-10-01
- Ben Recht: utility maximization is inescapable—we must learn when it's misapplied — beenwrekt · 2026-10-01
- Economist Paul Novosad: AI slop is crowding out authentic work — we need a firewall — paulnovosad · 2026-10-01