The White House is hosting a meeting with major tech companies, including Meta and Anthropic, to discuss a voluntary AI security framework. The plan aims to allow the government early access to advanced AI models to proactively test for vulnerabilities and hacking risks. Reuters reports that the finalized framework focuses on cybersecurity assessments of leading American AI technologies. This initiative will allow for evaluation of the models' susceptibility to attacks before potential misuse. The administration hopes this collaboration will strengthen national security in the rapidly evolving landscape of artificial intelligence. The tests will be voluntary, fostering cooperation between the public and private sectors. The move signals a growing concern over potential security threats posed by increasingly sophisticated AI systems.