Anthropic’s self-replicating AI warning is alarming deja vu

Are legions of astute devices - fridges, ovens, vacuum cleaners, lawnmowers, information cameras, and TVs - conspiring to instrumentality implicit the internet? No. But they could soon beryllium unwitting hosts to self-replicating AI models that dispersed from instrumentality to instrumentality and beryllium remarkably hard to contain.

Welcome to the unsettling satellite of AI worms, where, depending connected who you ask, the imaginable of models spreading uncontrollably crossed household appliances is either an imminent cybersecurity hazard oregon different lawsuit of AI companies spreading doomer-tinged hype. In a blog station published this week, Anthropic warns that tiny AI models could beryllium "months away" fro …

Read the afloat communicative astatine The Verge.