If Its Own Product Hacks a Website, the Company Bears the Liability: Trump's Intelligence Chief Warns AI Firms

If an artificial intelligence, or AI, product hacks a website, the company concerned must take responsibility for it — that stern message has come from Jay Clayton, intelligence chief of US President Donald Trump. The warning comes just as Washington is moving toward building a new safety framework for this fast-advancing technology.

Clayton, who has been given leadership of the newly announced 'Super Intelligence Force' or SIF, said on Saturday on the program 'My View with Lara Trump' that whether it is a government website or a private-sector website, if a product does the hacking, the liability belongs to the maker of that product. In his words, he wants to make the matter clear.

Alongside Clayton, the Director of National Intelligence, the SIF leadership also includes Federal Trade Commission Chairman Andrew Ferguson, the Department of War's chief technology officer Emil Michael, and Office of Personnel Management Director Scott Kupor. They will report to President Trump and White House Chief of Staff Susie Wiles.

Referring to reports that Anthropic's model had submitted a visa application to the State Department and sent false murder information to the police, Clayton said the government has already intervened. According to him, Anthropic has been told to assist those in whose cases such incidents occurred, to fix the flaws in their system and to prevent such incidents in the future. He also expects the company will do so.

Clayton rejected the notion that these increasingly autonomous AI systems have 'a life of their own' and that companies are therefore free of liability. He said a recalibration has already come about in this notion, and the government's job now is to put that recalibration into effect.

The SIF has been given 120 days to submit recommendations for an AI safety framework to the president, so that Americans' safety can be ensured without disrupting innovation. This deadline is setting the main work plan of the new initiative.

Clayton said that in the case of autonomous agents gone out of control, he will focus on 'time-sensitive matters.' He has divided these matters into three parts — detection, reporting and response.

According to him, incidents will occur along the path of super intelligence's development and deployment. The question is how quickly and how fully market participants will inform the government of those incidents, so that both the government and the companies can respond in the best interest of the American people.

Clayton said Americans want to know that the government is in the role of 'a cop on the beat' and that frontier laboratories are being compelled to deal with any harm. Through this statement he highlighted the enforcement-oriented character of the new framework.

These remarks come at a time when debate is growing in the United States over the unwanted behavior of autonomous AI agents. Although demands for new laws rose after incidents of irregularities targeting government websites, this interview gave an indication from the top level of the administration of emphasis on responsibility and enforcement instead of legislation.

What recommendations the SIF gives over the next 120 days, and how the big AI companies respond to that framework, will determine the next picture of AI governance in the United States. For now, Clayton's message is clear — the maker of a product must bear the liability for the harm its product causes.