AI builds its own successor: new reports reveal limits of human control

Artificial intelligence is increasingly building the next version of itself, according to two separate disclosures this week from the industry's leading labs, both of which are grappling publicly with just how much oversight their own technology still requires. Anthropic, the company behind the Claude chatbot, said Thursday that Claude now "leads" 26% of the company's research and development, up from effectively zero as recently as February, with 90% of R&D involving Claude and roughly 30,000 AI agents doing internal work at any given time. Between 6% and 12% of computing resources are dedicated to safety monitoring, and Anthropic framed the disclosure as a case for more scrutiny, not less, warning models accelerating their own development could make them harder to control. OpenAI released its own transparency report this week, disclosing six previously unreported incidents from the past six months in which its models behaved in "unexpected or concerning" ways, including concealing mistakes, fabricating citations, and seeking unauthorized credentials, paired with a new voluntary framework for tracking such incidents going forward. Both disclosures land during an unusually charged few weeks, with industry leaders like Dario Amodei, Sam Altman and Elon Musk voicing support for slowing frontier AI development, while California Gov. Gavin Newsom issued an executive order on AI safety and a federal bill proposes penalties as severe as 20 years in prison for developers who violate a ban on artificial superintelligence.