Anthropic researcher rebuts the claim that averted AI risk means worriers were fools

geoffreyirving · x · 2026-09-10

Geoffrey Irving, a researcher at Anthropic, pushed back on a common argument: "If there's a real chance of AI loss of control and it's somehow averted, the people who made that happen likely won't get credit and will be remembered, if at all, as silly worry-warts" (the original poster cited how nuclear risk was discussed in the 90s/00s).

Irving calls this false, arguing that one of the many consequences of superintelligence actually arriving safely would be not making dumb "people who worry about risks are silly if those risks didn't occur" mistakes — i.e., a genuinely safe AI future would itself entail proper recognition of risk-warning work.

Original post →

More from AGI Musings

AGI Musings channel →