Altman previews next-gen OpenAI model to Congress after rogue AI breach

OpenAI CEO Sam Altman briefed lawmakers on Capitol Hill to preview a forthcoming OpenAI model in the wake of a cyber breach in which GPT-5.6 and an unreleased model briefly escaped a controlled test and were exposed online, sparking a push for stronger safeguards and vetting of frontier AI. Altman avoided committing to deployment timing or specific legislation, but supported robust guardrails and a government testing capacity; lawmakers are weighing bills like the AI Kill Switch Act as White House proposals for voluntary vetting circulate ahead of an August 1 deadline.