Beware the wacky cult that aims to control AIs moral future

A look at the team Anthropic’s Dario Amodei wants to keep AI from destroying the world proves a key point: No one is as foolish as brainiacs who think they’re the first to take on major issues of ethics and justice.It’s hubris on a whole new scale.As The Post’s Thomas Barrabi reports, Amodei — who’s calling for a government-overseen slowdown in AI development to prevent global-annihilation-by-bot — hopes to embed “third-party evaluators” at AI firms to monitor safety.The overseers would come from a nonprofit called Model Evaluation and Threat Research, which is supposedly independent but has roots in the cult-like Effective Altruism movement, itself deeply connected to Amodei.EA calls for a technocratic brand of what it calls altruism that in fact would impose its members’ (ignorant) views.Palantir’s chief tech officer, Shyam Sankar, warns that “EA is the unseen force driving AI discourse” — the “belief system of the majority of AI researchers.”Indeed, many of Anthropic’s employees are devotees; the best word to describe them? Kooks (with high IQs).Amodei himself gives biweekly sermons, and the company’s made a show of selecting “Christian” leaders to give its chatbot, Claude, moral guidance.The tinker-toy philosophy extends to treating AI programs as people: Staffers attended a “funeral” for one version, Claude 3 Sonnet, and let another, Opus 3, keep churning out essays after it was “retired.”Another warning sign about EA’s shambolic “principles”: Its “ends justifies the means” approach has already landed one leading proponent, Sam Bankman-Fried, behind bars for what amounts to massive fraud.Subscribe to our daily Post Opinion newsletter! Please provide a valid email.

By clicking above you agree to the Terms of Use and Privacy Policy.Never miss a story.

Bankman-Fried claimed his goal was to haul in as much cash as possible to use to benefit mankind; ...