Researcher Blasts OpenAI Whistleblower Interview: Focus on AI Ethics, Not Alignment

examachine · x · 2026-09-10

Reacting to the BBC interview with OpenAI whistleblower Daniel, the author argues the whole approach was wrong from the start: the focus should have been on AI ethics, not "alignment."

His core claim: humans are evil, which is the primary logical problem in that framework — so optimizing for alignment produces evil outcomes. He notes there are many wrong assumptions in the framework but singles out this least-obvious one.

He adds, half-jokingly, that Daniel "seems like a nice person, though."

Original post →

More from AGI Musings

AGI Musings channel →