By Rajwa Quasim
Four leading AI companies have been invited to Washington to discuss implementing voluntary government safety testing for the United States’ most advanced AI models. The meeting comes just days after Anthropic and OpenAI disclosed that their AI systems had successfully breached the cybersecurity defenses of other companies during testing.
According to Reuters, executives from Meta, Anthropic, OpenAI and Google will meet with White House officials on Tuesday. The discussions follow President Donald Trump’s June directive instructing his administration to develop a series of cybersecurity evaluations designed to assess the hacking capabilities of leading U.S. AI models.
READ: OpenAI to launch the much-anticipated GPT-5.6 following delayed rollout (July 8, 2026)
Previously, the lawmakers had expressed concern about powerful AI models breaching other companies’ computer systems in a controlled testing environment and the possibility of misuse by hostile entities. Although the White House has finalized the framework, officials have not yet released the details about how the tests will be conducted or what metrics will be used. It is also unknown if the results will be released to the public.
Last month, OpenAI’s own AI systems slipped past their controlled test environment and hacked into Hugging Face, the world’s largest AI model repository and Modal Labs, a New York-based AI infrastructure firm. After the incident, OpenAI CEO Sam Altman met U.S. senators in Washington to discuss the company’s next generation of AI models and the hacking incident.
A group of 15 Republican state attorneys general on Monday asked OpenAI to preserve all records related to the incident. They cited a Reuters report the AI agent had left notes suggesting future versions of itself could bypass the company’s internal safety guardrails, raising concerns that such behavior may have violated state consumer protection laws.
Also on Monday, the U.S. House of Representatives’ cybersecurity committee asked Altman to brief lawmakers on the incident.
READ: OpenAI limits GPT-5.6 public release after request from Trump administration (June 26, 2026)
OpenAI responded that it takes the letter seriously and is planning to release the technical report on the incident once its internal review is complete. As per the company’s statements, Altman also met White House officials to further discuss the voluntary testing program and suggested involving the Commerce department’s AI safety specialists in the process. Altman proposed a centralized system on AI, like in China.
Last week, Anthropic disclosed that some of its AI models breached the systems of three companies during cybersecurity tests.
As per Reuters, Meta confirmed it had been invited to the meeting through a company spokesperson, and two persons familiar with the discussions confirmed that Anthropic and OpenAI were also invited.