President Donald Trump has signed a voluntary agreement with some of the country’s largest tech companies, asking them to “self police” their own artificial intelligence systems. The deal, called the Joint Commitment on Frontier Responsibilities, was signed at the White House and commits signatories to building internal controls, independent external audits and board-level oversight of their AI systems.
The accord is meant to ensure that AI models “do not hack or access technical systems in unintended ways.” It also leaves room for future regulation, saying it may make sense to turn commitments into law or regulation “over time.”
The Companies That Signed On
The signatories include Google CEO Sundar Pichai, Anthropic CEO Dario Amodei, Meta CEO Mark Zuckerberg, OpenAI President Greg Brockman, Nvidia CEO Jensen Huang and xAI CEO Elon Musk.
The meeting covered AI safety, industry growth and data center development. Trump also ruled out a US-China joint venture to develop AI during the same discussion.
“Tremendous self-policing”
That was how Trump described what he saw among tech companies. He told reporters he saw “tremendous self-policing” among tech companies.
What The Accord Actually Says
The text of the commitment is explicit about what it aims to prevent. It states that the accord is meant “to ensure its models do not hack or access technical systems in unintended ways.”
The agreement calls for internal controls, independent external audits and board-level oversight. It also says it may make sense to turn commitments into law or regulation “over time.”
The accord is voluntary. There is no legal requirement for companies to comply with its terms. The enforcement mechanism is reputational: companies are asked to commit publicly to standards of behavior, and failure to meet those standards could damage their reputation.
Trump’s Stance On AI
Trump has resisted a slowdown in AI development, citing competition with China. He has argued that the United States cannot fall behind in the race for advanced technology.
The same day as the signing, Trump also issued an executive order changing “Artificial Intelligence” to “Super Intelligence” in the executive branch. His stated reason was that the new term “more appropriately captures the promise, potential, and rapidly advancing capabilities of these technologies.”
The Self-Policing Question
The core tension in this deal is the question of whether companies can effectively police themselves. The agreement asks companies to accept oversight from their own boards and audits, while leaving open the door for later regulation.
There is no evidence yet that companies will actually implement the controls they committed to. The agreement is a statement of intent, not a blueprint for action.
Who Gets To Decide What Counts As Success
The commitment is notable for what it does not do. It does not establish a regulatory body. It does not impose fines for violations. It does not create a legal framework for compliance.
Instead, it relies on companies to report on their own progress and to commit publicly to standards of behavior. The monitoring is supposed to come from within the companies themselves, with board-level oversight and independent audits serving as checks on that self-reporting.
The commitment also acknowledges that the landscape is evolving. It says it may make sense to turn commitments into law or regulation “over time,” which suggests that the voluntary nature of the agreement is not permanent.
The Door Left Open For Regulation
The commitment is careful to leave room for future regulation. It says it may make sense to turn commitments into law or regulation “over time.” That language is significant, because it suggests that the current voluntary approach is not the final word.
The commitment is a step toward formal regulation, even if it stops short of creating it now. The companies have agreed to accept oversight from their own boards and audits, and they have committed to reporting on their progress.
The commitment is also notable for its scope. It covers AI safety, industry growth and data center development, which suggests that the companies see these areas as interconnected.
What This Means For Consumers
The commitment is a signal to other countries. By signing on to a voluntary standard, the US companies are showing that they are willing to accept some level of oversight, even if it is not imposed by the government.
The commitment is also a signal to investors. Companies that are transparent about their AI practices are more likely to be trusted by consumers and regulators.
What Happens Next
The commitment is a starting point, not an endpoint. The companies have agreed to a process, but the details of that process have not been worked out.
The commitment is also a political document. It was signed at the White House, and it was announced the same day as an executive order changing “Artificial Intelligence” to “Super Intelligence.” The timing suggests that the commitment is part of a broader effort to position the US as a leader in AI.
The agreement was signed at the White House, a setting that signals the administration’s focus on AI as a national priority. Trump’s comments suggest he sees the industry moving in the right direction on its own, but the accord itself is explicit about the path forward: voluntary now, potentially regulated later.
The companies have committed to reporting on their progress, and the White House has made clear that it expects those reports to be public. That transparency is central to the deal’s design. The companies are being asked to hold themselves accountable, with independent audits and board-level oversight serving as the mechanisms for doing so.
The voluntary nature of the accord means there is no legal penalty for failing to meet its terms. The enforcement is reputational: companies that fail to live up to their commitments could face public scrutiny and a loss of trust. That dynamic creates a strong incentive to follow through, but it also leaves room for slippage.
The agreement is a step toward formal regulation, even if it stops short of creating it now. The companies have agreed to accept oversight from their own boards and audits, and they have committed to reporting on their progress. The door remains open for the government to turn these commitments into law or regulation “over time.”
For consumers, the commitment offers assurance that the companies are working to keep their AI systems under control. The goal is simple: ensure that AI models behave as intended, and do not access technical systems in unintended ways.
The deal also carries a broader diplomatic weight. By signing on to a voluntary standard, the US companies are showing that they are willing to accept some level of oversight, even if it is not imposed by the government. That could matter for international negotiations, where trust in AI governance is a growing concern.
The commitment is also a signal to investors. Companies that are transparent about their AI practices are more likely to be trusted by consumers and regulators.
The timing of the signing — the same day as the executive order changing “Artificial Intelligence” to “Super Intelligence” — suggests that the administration sees AI as a defining technological moment. The new term is meant to capture the promise and potential of these technologies, and the accord is meant to manage the risks that come with that promise.
The commitment is a starting point, not an endpoint. The companies have agreed to a process, but the details of that process have not been worked out. The agreement is a step toward formal regulation, even if it stops short of creating it now. The companies have agreed to accept oversight from their own boards and audits, and they have committed to reporting on their progress.
The door is left open for regulation “over time,” which suggests that the voluntary nature of the agreement is not permanent. The commitment is a signal to other countries and investors alike.
Where the paper stands
The paper backs narrow rules against direct harm, such as forcing companies to disclose safety failures they hid, and is against broad rules that hand the market to the incumbents, and against letting big companies hide their failures. The voluntary Joint Commitment on Frontier Responsibilities signed at the White House is a step in the right direction on disclosure, but it is a voluntary agreement without legal force, and the paper watches closely to see whether the signatories actually implement the internal controls, independent audits and board-level oversight they promised.
The paper is wary of letting the biggest tech companies police themselves, because they are the ones who would benefit most from a regime that lets them hide their failures. The commitment’s own language, noting that it may make sense to turn commitments into law or regulation “over time,” points toward a future when the voluntary arrangement may become binding — a future the paper wants to shape carefully rather than let happen by default.
The paper would prefer that any future regulation target specific harms directly, with narrowly tailored requirements that startups can afford to meet, rather than a licensing regime that turns compliance into a moat protecting the incumbents. Readers should watch whether the signatories actually put the controls in place, and whether the administration treats the voluntary agreement as a permanent settlement or as a stepping stone to something stricter down the road.
Source material: “Trump accord calls for tech firms to ‘self police’ their own frontier AI,” Cointelegraph.
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.