The United States finalises voluntary AI safety tests, according to a White House official cited by Channel NewsAsia. The announcement follows discussions involving leading AI companies, including meetings with OpenAI’s leadership. The Independent reports that OpenAI CEO Sam Altman visited the White House last week to discuss the voluntary tests and OpenAI’s upcoming AI models. Both accounts frame the tests as part of a US effort to encourage responsible development of advanced AI systems without imposing binding requirements. The Independent also characterises the situation as involving breaches by major companies, in the context of the broader rollout of the testing framework. However, the specific details of what constitutes a breach and which obligations were not met are not provided in the excerpts available here. Overall, the reporting indicates that the US is moving from planning to completion of the voluntary safety testing process, while engaging major AI developers ahead of the release of new model versions. The statements emphasise the voluntary nature of the testing and the role of industry engagement in shaping the approach.