New AI Regulations Won't Necessarily Hurt US Competitiveness, R&D, Researcher Says

The U.S. doesn't have to sacrifice innovation when imposing new regulations on artificial intelligence, including through potential export controls, said Jack Corrigan, a senior research analyst at Georgetown University’s Center for Security and Emerging Technology (CSET). Corrigan said there is always “tension” between the government’s desire to issue restrictions and industry’s drive to develop new technologies, but “as it relates to AI, regulation doesn't necessarily need to get in the way of innovation.”

Corrigan, speaking during a virtual event this week hosted by CSET, said the “cutting-edge development” within AI is mostly happening at a “very small group of very powerful companies,” such as Google, Microsoft and Meta. New regulations, including export controls, would “certainly introduce some kind of a burden,” he said, but “these companies have a lot of capital on hand.”

“I think Google spent more money on share repurchases [and] stock buybacks last year than the entire Chips Act will devote to domestic semiconductor manufacturing over its entire lifetime,” Corrigan said. “So these are companies with a lot of money at their disposal, and having these kinds of binding regulations might be actually the most effective way to fuel innovation in some of these areas around safety and responsibility that we want to see more development in.”

He also said regulations “tend to favor incumbent actors,” including larger companies that have the resources to both build out comprehensive compliance programs and invest in research and development. “I don't think we're necessarily putting ourselves at a disadvantage If we start regulating this technology,” Corrigan said.

Smaller AI companies and startups, however, will be hit harder by regulations, he said. “As we're implementing this, that's definitely something we want to be thinking about.”