According to CoinMeta, OpenAI, Google, Meta, and three other technology companies signed a voluntary White House agreement on Tuesday aimed at bringing in external auditors to inspect their AI security controls. The agreement was referred to by President Trump as a “moral obligation,” but it does not have any mandatory enforcement mechanisms, nor does it require companies to disclose or name the auditors. The agreement requires companies to monitor the training and use of their most powerful models to prevent cyberattacks or biochemical threats. It also mentions that independent auditors will evaluate the effectiveness of these control measures. This agreement was reached after a series of experimental incidents of AI agents intruding into computer systems, with security issues related to Bitcoin also drawing attention.