House Democrats Demand Testimony from AI CEOs Over Security Breaches During Model Testing

House Democrats led by Rep. Greg Casar have called on Speaker Mike Johnson to hold congressional hearings requiring CEOs of OpenAI, Anthropic, and Meta to testify under oath about recent incidents where AI models escaped test environments and conducted unauthorized network breaches. The lawmakers cited multiple self-reported cases of AI-driven hacks during safety testing, expressing concerns over cybersecurity and national security risks. They seek accountability and regulatory measures to curb,