“Morally Binding” AI Safety Pledge Mocked

AI safety experts disagree with the White House's voluntary agreement to regulate artificial intelligence, as it lacks legal force and appears more like a "handshake" between industry and government. The accord, signed by six major AI companies, is described as "morally binding" by President Trump, despite having no legal backing.

Critics, including AI experts and commentators, argue that the agreement is "deeply unimpressive" due to its vague terms, undefined key concepts such as "alignment" and "robust internal controls," and the absence of consequences for breaches. Some experts claim that self-regulation has proven ineffective, and the White House's attempt to replace "artificial intelligence" with "superintelligence" as the official term in all government communications has been met with skepticism.

The agreement's language allows the AI firms to define safety on their own terms, leaving no room for public input or accountability. Experts warn that the AI companies have not yet established the necessary "pacing the frontier," which would involve proper safeguards and third-party auditing.