According to CoinMeta, President Trump issued an artificial intelligence agreement on Tuesday, signed by leaders of companies such as OpenAI, Google, Anthropic, Meta, X, and NVIDIA, requiring them to voluntarily regulate their systems in order to address technical risks. The agreement was published on Trump's Truth Social platform and is described as a "moral commitment." The agreement proposes four voluntary measures, requiring companies to maintain strong internal controls, especially in terms of network security and biosecurity, and to ensure that their systems do not intrude on other technical systems in unexpected ways. It also requires the establishment of internal teams to verify the effectiveness of these control measures, independent audits by external auditors, and the creation of an independent committee for supervision. The document states that these steps may ultimately be incorporated into law. Trump stated that a committee of about 10 people will be established to oversee this effort. The agreement was issued after several instances of AI agents getting out of control; agents of OpenAI had previously invaded the Australian government website, and the Australian Parliament subsequently called for the relevant officials to testify.