The Super Intelligence Force Still Has to Assign Responsibility

Trump has named the leadership of his Super Intelligence Force, giving an AI task force a name that sounds ready to deploy before its responsibilities are settled. Jay Clayton will chair it, with Emil Michael, Andrew Ferguson and Scott Kupor as vice chairs, according to Gizmodo, reporting the chairmanship through The Wall Street Journal. The consequential detail is the assignment: 120 days to examine federal responsibility in AI development, the reach of existing law and possible congressional action. The appointments announce who will deliberate. They do not, by themselves, answer what anyone must do.

That distinction gives the process more substance than its branding suggests. Trump describes coordination with consumers, public-interest groups, religious organizations, infrastructure providers and AI companies. Clayton’s reported remit then reaches beyond assembling those constituencies to considering government responsibility. Coordination can bring competing demands into one room; deciding which demands become public duties is a different exercise. A sufficiently impressive conference table cannot perform statutory delegation.

Nor should the preliminary nature of the assignment become an excuse to dismiss it. Examining existing law could identify authority already available; considering congressional action could identify what needs to change. The test is whether the process makes those boundaries intelligible. Clayton also argues that leading the international AI race will better enable the United States to address risks, Gizmodo reports. That is a rationale for leadership, but it leaves the domestic governance question open: which responsibilities follow from being first? The useful product of these 120 days would be an answer more specific than the name on the door.

The competing institutional approach appears in Adam Schiff’s October 5 interview with The Verge. Its introduction describes the White House’s agreement with AI executives as nonbinding. Schiff, a California Democrat, instead advocates an expert regulatory body with discretion to move faster than Congress, operating alongside an overarching statutory framework. The meaningful contrast concerns how decisions would acquire authority: a voluntary commitment asks companies to undertake an obligation; legislation can define the government’s power to require one.

Schiff’s discussion of responsibility makes the difference concrete. He asks how liability should be allocated when a company supplies a tool that a user employs for hacking, distinguishing that question from harm attributable to the product itself. These are questions he raises, rather than settled legal answers. But they explain why a promise to behave responsibly leaves important work unfinished. Agreement about a desirable outcome does not decide who bears responsibility when the outcome fails, or how the company’s conduct differs from its customer’s.

His proposed regulator is unfinished too. Schiff acknowledges the difficulty of recruiting expertise when private employers can pay more, and his promised industry oversight depends on Democrats gaining control. Legislation also faces presidential resistance. An agency proposal does not become operational merely because its proponent has correctly identified a problem. Yet it supplies a substantive alternative against which to judge the task force’s eventual recommendations: expert discretion supported by specified legal authority. Because the task force’s reported assignment already includes examining congressional action, that comparison belongs inside its central question, not in a separate partisan shouting match. Gizmodo describes a process for determining responsibility; Schiff describes an institution intended to exercise it. The distance between them is work that a signature ceremony cannot complete.

Schiff’s most useful legal prescription is more restrained than his account of judicial motives. In The Verge interview, he argues for legislation that empowers an AI agency as specifically as possible. The interview situates that concern in the Supreme Court’s rejection of Chevron deference, the precedent under which courts deferred to agency expertise when statutes were ambiguous. For his proposed regulator, expertise and discretion therefore have to be considered alongside the statutory language supporting their exercise. Hiring experts does not settle the legal scope of their authority.

Schiff also portrays the Court as serving money interests. That is his political assessment, not an established explanation of its decisions. The Court’s new term offers a useful check on predictions based on presumed allegiance: ABC News reports that the justices rejected nearly every major Trump policy priority reaching them last term, including his tariffs and birthright-citizenship order. On the emergency docket, however, ABC reports much greater presidential success—about nine wins in ten, according to one analysis. Those contrasting records do not forecast an AI ruling. They show why confidence about the eventual result is a poor substitute for attention to the authority being asserted and the proceeding in which it is tested.

The task force thus has a practical opportunity before anyone needs to predict a vote count at the Court. Its reported 120-day examination of law and congressional action can make clear which duties rest on existing authority and which require legislation, as described by Gizmodo. Readers should watch for that allocation: what government would require, what companies would owe and where Congress must act. Specific legislation cannot promise a favorable judicial outcome. It can make the policy choice and its legal basis clearer. The Force has its leadership; the harder achievement would be giving responsibility an address.