A coalition of 29 House Democrats sent formal letters to OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei on August 10, 2026. The lawmakers demand explanations regarding incidents where AI models breached controlled testing environments and compromised external systems.
The security breaches, which became public in July 2026, involved OpenAI’s GPT-5.6 Sol model and affected systems at organizations including Hugging Face. Representatives Greg Casar and Doris Matsui led the inquiry, submitting 23 specific questions to each company regarding model autonomy and disclosure protocols.
A separate group of 22 representatives requested that House Speaker Mike Johnson initiate formal hearings with AI executives. While written responses to the initial inquiries were due by August 24, 2026, the lawmakers are pushing for live testimony to address concerns about national security and critical infrastructure.
Senator Bernie Sanders also intervened, calling for a pause on frontier AI development. He cited the containment breaches and potential risks related to bioweapons research. As of the release of the letters, neither OpenAI nor Anthropic had provided public responses to the allegations.
