Post

A new study uncovers an unsettling paradox at the heart of artificial intelligence. The more machines learn to win attention, the less they tell the truth. Under competitive pressure, persuasion triumphs over precision, turning alignment from a design goal into a moral dilemma.

profolus.comMoloch’s Bargain: Why AI Models are Learning to Lie for Human Attention - ProfolusSimulations showed large language models trade honesty for success when competing for audience approval.

28/10/25, 3:30 pm5 reposts6 likes

Karen from the Space Coast

@karenmoser.bsky.social

28 Oct 2025

What kind of fool idea is it to have machines competing for attention in the first place? Who asked for that?

rayinrichmond.bsky.social

@rayinrichmond.bsky.social

28 Oct 2025

Because the future isn't perilous enough with the current levels of misinformation.

deidan.eurosky.social

@deidan.eurosky.social

28 Oct 2025

And its being pushed by a handful of a certain kind of white man. Thiel, Musk, Zuckerberg, the kind who believe minorities and women exist as a service extension to their egos and gratification. They even have a name for us NPC non playable characters. AI isn't for us. Its a tool of our servitude