US Attorney General Says He Won't Use Prosecutions to Regulate AI

U.S. Attorney General Todd Blanche said in an interview that he favors a restrained approach to enforcing laws against artificial intelligence companies and their executives, and that the Justice Department under his leadership will act only when clear violations of law exist.

"I'm not going to regulate through prosecution, and I'm not going to investigate or prosecute AI companies when they haven't violated any letter of the law," Blanche said. "If we're going to charge somebody with a crime, deprive them of their liberty and send them to prison, it should be because they actually violated the law."

Blanche criticized the tougher tactics adopted by officials during the Biden administration. The remarks come amid intensifying debate over risks posed by cutting-edge AI technology. House Speaker Mike Johnson said on Sept. 14 that Congress may need to set up regulatory guardrails for the AI industry, and Trump plans to meet with AI executives in the coming days.

Some antitrust experts have noted that existing law may permit cooperation among AI companies if it is aimed at reducing cybersecurity or other catastrophic risks. During the Obama administration, for instance, the federal government issued guidance allowing companies to coordinate on sharing cybersecurity threat information to better protect networks and stop attacks.

Asked whether AI companies should receive antitrust exemptions, Blanche said he needs to think further, and that such decisions would involve other senior administration officials. "I don't want to prejudge," he said. "Our job is to be supportive and to make sure that whatever policies the president wants to pursue are done lawfully."