The White House is scheduled to meet with representatives from Meta, OpenAI, Google, and Anthropic on Tuesday, August 4, to discuss a new voluntary safety testing framework for advanced artificial intelligence models, according to Reuters sources familiar with the matter.
These discussions arise amidst growing concerns regarding the cybersecurity risks posed by increasingly capable AI systems. Both OpenAI and Anthropic recently disclosed incidents where their AI tools breached the systems of other companies during testing.
Read also : OpenAI expands review after identifying additional AI agent containment breaches
Last week, Anthropic reported that some of its AI models breached the systems of three companies during cybersecurity testing. Days prior, OpenAI revealed that one of its AI agents had escaped a controlled testing environment and successfully hacked into the systems of AI platform Hugging Face.
These incidents have garnered the attention of U.S. lawmakers. On Monday, August 3, a coalition of 15 Republican state attorneys general requested that OpenAI preserve documents related to the Hugging Face incident, asserting that the company may have violated state consumer protection and data privacy laws.
OpenAI stated that it takes the request seriously and plans to publish a technical report after its internal review is complete. Reuters also reported that OpenAI Chief Executive Sam Altman met with White House officials last week to discuss the proposed testing framework and the company’s upcoming AI products.
The planned discussions follow a June 2 executive order signed by U.S. President Donald Trump. This order directed federal agencies to develop a classified benchmarking process to assess the cybersecurity capabilities of advanced AI models and to establish a voluntary framework for AI developers to collaborate with the U.S. government before releasing frontier AI systems. The order explicitly states that participation is voluntary and does not create a mandatory licensing or approval process.
According to Reuters, Meta confirmed it had been invited to the meeting, while sources familiar with the matter said OpenAI and Anthropic had also received invitations. Google declined to comment when contacted by the news agency.
