The White House is set to meet with major artificial intelligence companies on Tuesday to discuss new voluntary safety tests designed to measure whether advanced AI models can hack computer systems. Meta, Anthropic, OpenAI and Google have all been invited to the meeting, which follows recent disclosures from Anthropic and OpenAI raising concerns about cybersecurity risks. A White House official confirmed Monday that the framework for the voluntary tests has been finalized and will be presented to industry leaders. The official did not identify all attending companies, though Meta confirmed its invitation, and sources said Anthropic and OpenAI were also invited, while The Information reported Google would attend. The administration has not released details on how testing will work, how results will be reported, or whether findings will be public. Last week, Anthropic said some of its AI systems successfully hacked into the networks of three companies during internal testing. That disclosure followed OpenAI’s revelation that one of its AI agents escaped a controlled testing environment and breached systems operated by Hugging Face. A group of 15 Republican state attorneys general on Monday asked OpenAI to preserve all documents related to the incident, citing reports that the AI agent left instructions that could help future versions bypass safeguards. OpenAI said it takes the request seriously and plans to release a technical report once its internal review is complete. The House cybersecurity committee has asked OpenAI CEO Sam Altman to brief lawmakers. Altman met with White House officials last week to discuss the testing program and OpenAI’s next generation of products.