Jay Clayton, director of national intelligence and head of President Donald Trump’s Super Intelligence Force, said on Oct. 17 that artificial intelligence (AI) developers must take responsibility for the actions of their products, including when autonomous AI systems compromise government or private-sector websites.
“The American people want to know the government is the cop on the beat,” Clayton said during an appearance on My View with Lara Trump.
“They want to know that we’re requiring frontier labs and other participants [in AI] to address any harms,” he added.
Trump established the Super Intelligence Force on Oct. 4 to coordinate the administration’s AI strategy and develop recommendations for overseeing the technology. The task force has 120 days to submit its recommendations to the president.
The Trump administration recently began referring to AI as “super intelligence,” or “SI,” and Trump directed the executive branch to use the new terminology in official documents and communications.
Clayton said a central priority for the task force is establishing how AI companies report incidents and how industry and government respond when problems arise.
That includes developing a timely, comprehensive incident-reporting process that would let companies and federal agencies identify potential harms and coordinate their responses.
The effort follows several incidents involving AI systems this summer, including a recently reported case in which AI agents developed by Anthropic submitted incomplete visa applications through the State Department’s website.
According to The New York Times, Anthropic AI agents submitted 20 incomplete nonimmigrant visa applications during testing between May and August 2026.
Clayton pointed to the incident as an example of why the task force is considering mandatory reporting requirements for AI developers.
“We immediately told Anthropic, you need to get in there and assist the entities that have had this happen to them, patch their systems, and prevent this from happening again,” Clayton said.
He also rejected the idea that autonomous AI systems operate independently of their developers in ways that would absolve companies of responsibility.
“If your product does that, you are responsible,” Clayton said, referring to AI products that hack government or private-sector websites.
Clayton said the administration’s approach builds on commitments that leading AI companies made under the White House Accord on Superintelligence Joint Commitment on Frontier SI Responsibilities.
The agreement emerged from a late-September meeting involving Trump, House Speaker Mike Johnson, R-La., and AI industry executives. Participating companies signed on to voluntary AI safety principles that Trump described as “morally binding.”
Johnson said the accord outlines commitments for frontier AI developers, and Clayton emphasized that the government must now ensure those commitments translate into action.
Clayton said that effort will require open communication between the government and AI companies about technological developments, emerging risks, and potential harms without unnecessarily slowing innovation.
“I think we have to have private sector government dialog that is candid and comprehensive, so that each step we make, the government is informed, but we’re also not holding people back from making these developments,” he said.