Trump and AI Chiefs Sign a Voluntary "Accord on Super Intelligence" With No Penalties for Breaking It

The signatories wrote: "We believe every company is responsible for developing its own technology safely and in a way that builds trust with customers and the public." CBS notes that the text itself concedes the limits: "Over time, it may make sense to codify these steps into laws or regulations."

Binding in spirit only

The accord is voluntary and carries no legal penalties for those who ignore it. Trump nonetheless described it in strong terms. "It's almost like a constitution, in a way," he said, adding: "I think it's morally binding." He also put the arrangement in terms of mutual pressure, and said he was seeing "tremendous self-policing". Zuckerberg called the steps "good steps forward" and the discussions "historic", and Huang said: "There's no conflict between innovation, technology and safety."

Renaming and a new committee

The event came with other announcements. Trump signed an executive order directing federal agencies to use the term "super intelligence", shortened to SI, in official documents, and said he plans to name an AI czar within days and form a 10-person oversight committee, according to CBS and SiliconANGLE.

Democrats were unimpressed by the vocabulary. Senator Mark Warner of Virginia said: "The companies building the most powerful AI systems are warning us that the technology is advancing faster than our safeguards," and the signing arrives as lawmakers weigh binding rules.

What to watch

A voluntary framework is only as strong as its enforcement. The commitments to external audits and independent board committees are the parts that could, if carried out, produce evidence that outsiders can check. Whether the czar, the committee and any later legislation turn that evidence into obligations is the open question, and the accord's own wording suggests even its signatories expect that debate to continue.