The AI Hoax Needs a Hotline

The administration’s AI posture has acquired the familiar shape of modern executive politics: deny the problem in public, build a bureaucracy around it in private, and call the contradiction leadership. Donald Trump reportedly dismissed fears about AI, alongside affordability concerns, as a hoax—while his Treasury secretary was simultaneously proposing an AI-incident notification mechanism to China for events that could affect national security. The political backlash at home and the diplomatic outreach abroad are not quite opposite policies. They are opposite registers of policy: campaign minimization and state-capacity realism.

Treasury Secretary Scott Bessent’s proposal is thinly described: no public detail on covered incidents, procedures, Chinese acceptance, or a signed agreement. Still, its premise is unmistakable. Two rival powers that see AI as strategically decisive may need a way to alert one another when a machine-generated event crosses into national-security territory. That is a modest proposition, not a treaty of digital Versailles. Yet it concedes what the domestic rhetoric evades: AI risk is not merely a talking point deployed by Democrats with too much time and a PDF reader.

The China talks also bundled AI with tariffs, trade and a proposed Board of Trade ahead of a Trump-Xi meeting. That is how power actually works: safety is negotiated beside market access, not in a purified seminar room where every participant has read the alignment literature. The administration wants American AI dominance and more transparency with China. Those ambitions may coexist, but they require rules, channels and people empowered to use them. An AI Force without an incident protocol is just a branding exercise with unusually expensive servers.

Republican donors are not objecting to technological acceleration on philosophical grounds. They are observing, with the apprehension of people who have seen a focus group, that voters can worry about more than one thing at once. The POLITICO reporting carried by Yahoo describes donors who thought Trump’s treatment of AI fears made him seem out of touch—particularly against higher mortgage rates, record diesel prices and gasoline averaging $4.47 a gallon, according to AAA.

The important political fact is not that AI has displaced affordability. It plainly has not. A Marquette University Law School poll cited in the article put approval of Trump’s handling of inflation and cost of living at 19 percent, and gas prices at 20 percent. Vice President JD Vance and House Minority Leader Hakeem Jeffries were both concentrating on cost-of-living questions in their respective party meetings. The gas pump remains a much more efficient political communications device than a frontier model’s risk assessment.

But the donor anxiety is rational. The article cites a POLITICO poll finding roughly two-thirds of Americans saw at least moderate risk that advanced AI could destroy humanity, with majorities across party, gender and age groups. One need not treat public fears as a technically calibrated forecast to recognize them as politically material. When the same government says ordinary anxieties are imaginary while organizing international guardrails for the underlying technology, voters may conclude—rather impolitely—that someone is being managed. And it is rarely the machine.

JB Pritzker’s response is dramatic in language and deliberately unglamorous in machinery. In an ABC News interview, the Illinois governor called AI “dangerous on the level of nuclear weapons,” urged federal and international action, and rejected both industry self-regulation and a blanket ban on superintelligence. That leaves him in the politically unfashionable middle: neither let the companies write the rules nor pretend the technology can be legislated back into nonexistence.

Illinois has already enacted a more concrete version of that instinct. The law Pritzker signed in July requires AI companies to submit annual third-party audits of their safety plans and disclose safety incidents to the state. The evidence supplied does not establish how the law will be enforced or whether it will work. But it does identify the basic constitutional and administrative question that Washington keeps trying to turn into content: if a company develops a system with broad social and security implications, who gets to inspect its risk claims, and what must it report when things go wrong?

That question links Pritzker’s audit model to Bessent’s proposed U.S.-China alert channel. One is domestic oversight; the other is interstate crisis management. Both rest on an almost embarrassingly old-fashioned proposition: power should create records, accountability and notice before it creates excuses. Trump’s preference for acceleration may be strategically coherent; slowing American development while China advances carries obvious risks. But acceleration without disclosure is not national strength. It is a procurement philosophy dressed as destiny—fast, confident and, when it fails, suddenly eager to find the hotline.