On September 29, leaders from six of the country’s most powerful AI companies stood at the White House and signed a one page promise to police themselves. CBS News reported that OpenAI, Anthropic, Google, Meta, Nvidia and Elon Musk’s SpaceXAI agreed to internal evaluations, outside audits and reviews by their own boards. Asked whether the accord bound anyone, Trump gave the honest answer: “I think it’s morally binding.” He also called it “almost like a constitution, in a way.”
Five days later, the government part arrived. CBS News reported that Trump created the Super Intelligence Force, led by Director of National Intelligence Jay Clayton with Federal Trade Commission chair Andrew Ferguson, Pentagon chief technology officer Emil Michael and Office of Personnel Management director Scott Kupor. The Associated Press confirmed that the group will report to Trump and White House chief of staff Susie Wiles.
The government’s answer to an industry asking for trust is a voluntary accord and a task force whose four named leaders all answer to the president. The public announcement named only administration officials and described no penalty or compulsory process. It promised engagement with consumers, public interest groups, religious organizations, infrastructure providers and industry leaders. Engagement is a meeting. Oversight requires authority.
A promise cannot do a law’s job
The accord’s structure contains every incentive problem regulation is supposed to solve. The companies run the internal evaluations. Outside firms audit them, but the published text leaves unanswered who selects or pays those firms. Each company’s board reviews the resulting reports. ABC News noted that the document says companies “believe” each signatory “should implement” the controls. The outlet said it was unclear whether the pledge required OpenAI or Anthropic to do anything beyond audits they had already accepted.
The accord concedes its own weakness. “Over time, it may make sense to codify these steps into laws or regulations,” it says. That sentence describes the missing part. Codification brings legal duties, public rules and consequences for breaking them. Trump offered moral obligation instead.
We have watched this incentive structure fail. The credit rating agencies that blessed dangerous securities before the 2008 financial crisis were paid by the issuers they rated. A Securities and Exchange Commission statement calls that model a fundamental conflict of interest and says understated risk helped build the leverage that fed the crash. A reviewer can be competent and sincere while the party being reviewed still controls the commercial relationship. AI audits need rules that survive bad incentives. Honor is a poor substitute.
Trump chose the intelligence chief
ODNI’s biography says Clayton is a lawyer who led the SEC during Trump’s first term and later served as the top federal prosecutor in Manhattan. The Senate confirmed him as Director of National Intelligence in July, and he was sworn in on August 3. He now heads both the Super Intelligence Force and an 18 organization intelligence community that includes the CIA, NSA, FBI and military intelligence services.
Sunday’s announcement discussed coordination and outreach. It never invoked Clayton’s intelligence authorities. The leadership choice still matters because it puts the country’s intelligence chief at the center of a civilian technology debate about jobs, privacy, discrimination and consumer safety. Michael’s Pentagon role adds another national security voice. Ferguson at least leads a consumer protection and competition agency. Kupor runs the federal workforce. The four named leaders omit independent scientists, worker representatives, civil liberties advocates and state regulators.
The task force will reportedly have 120 days to produce a report. TechCrunch, citing the Wall Street Journal, says its charter calls for plans to respond to AI threats while preventing “overregulation and regulatory capture.” The capture warning is an insult. The regulated industry just wrote the voluntary standards.
Speed needs a public brake
Congress has yet to pass a comprehensive AI law. China is racing toward the same capabilities. A public commitment by six competing companies can create audit practices now, while legislation takes months or years. Nvidia chief Jensen Huang argued at the White House that innovation and safety can advance together. He is right about that principle.
The principle cuts against Trump’s approach. Without enforcement, safety remains a corporate program. Public protection requires law. Axios reported that even Senate Majority Leader John Thune wants Congress to codify safety protections. Republican Sen. John Hoeven put the point more bluntly: “You have to have the strength of statute and law and regulation.” They are hardly radicals trying to smother technology.
Mandatory testing allows research to continue. Incident reporting leaves America’s competitive position intact. Public rules for independent audits are compatible with innovation. These are ordinary demands placed on industries whose failures can hurt people who never bought their products and never consented to the risk.
The rebrand covers an empty rulebook
The branding exercise exposes the administration’s priorities. On the day the accord was signed, Trump issued Executive Order 14434, directing the executive branch to replace “artificial intelligence” with “Super Intelligence” in official communications. The order changed the vocabulary and left safety rules untouched.
In a statement carried by CBS News, Sen. Mark Warner summarized the arrangement: “The companies building the most powerful AI systems are warning us that the technology is advancing faster than our safeguards. The president’s response? To rename it and tell the companies developing it to regulate themselves.”
The new task force deepens the problem Warner named. It concentrates AI policy in four administration officials who report to Trump and Wiles, with the intelligence chief at the top. Consumers and public interest groups may be invited to talk. The companies retain the power to evaluate themselves, present audits to their own boards and decide whether a moral obligation is enough.
Trump compared the accord to a constitution. The comparison works if the constitution was written by a king. The companies make the rules, the president appoints the referees, and the public gets no enforceable rights at all.