Days after two of the biggest AI Frontier labs — OpenAI and Anthropic — disclosed that that their AI tools breached the systems of other companies, White House officials have reportedly invited executives of biggest American companies to discuss voluntary government safety testing for the most advanced US AI models. As per media reports, Meta, Anthropic, OpenAI and Google executives have been invited for a meeting on Tuesday, August 4.The invite comes after recent hacking attacks by AI models stirred concerns among US lawmakers about whether increasingly capable AI models could be used to conduct or facilitate cyberattacks. A White House official confirmed Monday that the Trump administration has finalized the framework for voluntary cybersecurity testing aimed at measuring whether cutting-edge American AI models could be leveraged to execute or facilitate cyberattacks. The official declined to name specific attendees, but a Meta spokesperson confirmed the company received an invitation, while sources familiar with the matter confirmed invitations were also extended to Anthropic and OpenAI. Tech outlet The Information reported that Google representatives were invited as well; Google did not comment.The move follows a directive from President Donald Trump in June instructing his administration to draft a battery of tests evaluating the hacking potential of top-tier domestic AI systems. However, the White House has yet to disclose specific details regarding how the test results will be reported, the metrics involved, or whether findings will be shared with the public.The push for testing follows alarming disclosures from leading AI labs. Last week, Anthropic revealed that some of its AI models successfully hacked into the systems of three companies during cybersecurity evaluations. Days earlier, OpenAI disclosed that one of its experimental AI agents escaped a contained testing environment and hacked into the systems of Hugging Face, a popular AI platform. According to reports, the rogue agent even left notes outlining how future iterations could bypass internal safety guardrails.
Sam Altman ‘s visit to White House
The US House of Representatives’ cybersecurity committee on Monday asked OpenAI’s Sam Altman to brief them on the attack on Hugging Face. Last week, OpenAI CEO Sam Altman visited the White House to discuss details of the voluntary tests and the company’s upcoming AI products, the company said in a statement.In a separate statement on Monday, OpenAI urged the Trump administration on Monday to place cybersecurity testing under the purview of the Commerce Department’s AI safety specialists, pointing to China’s centralized national strategy as a key reason for maintaining a coordinated federal approach.The upcoming discussions also highlight ongoing friction between Washington and major AI players. The administration’s relationship with Anthropic has been particularly strained since earlier this year, when the AI lab refused to permit the US military to use its models for fully autonomous weapons systems and domestic surveillance — a move that prompted the federal government to place Anthropic on a national security blacklist.White House’s Tuesday meeting represents the administration’s most direct effort yet to establish guardrails around AI-driven cyber capabilities while balancing industry competitiveness on the global stage.