ChatGPT Companies Want Regulation: Califf

FDA commissioner Robert Califf says the agency has met with officials of the leading ChatGPT companies “and they want to be regulated. But even they don’t really have good suggestions now on how to do it.” Califf commented at a recent Biotechnology Innovation Organization convention, as reported by Bloomberg Law.

The report says Califf often refers to large language models popularized by ChatGPT in many presentations to push the idea that he wants FDA to get ahead of regulating the disruptive technology. (Large language models (LLMs) emerged around 2018 and do not yet have a formal definition but are often referred to as deep learning models with millions or even billions of parameters, that have been “pre-trained” on a large body of materials. LLMs are said to be general purpose models that excel at a wide range of tasks, as opposed to being trained for one specific task.)

Large language models “are going to be transformative and it’s got to be regulated,” Califf said at the convention.

Bloomberg Law says it’s been told by FDA attorneys, regulatory experts, and legal scholars that the agency has the power to regulate the use of large language models in the products it regulates. It says the agency’s authority will largely depend on whether the intended use of the models is to treat or diagnose a condition or disease.

The article quotes attorney David Peloquin (Ropes & Gray) as saying that some of the principles that are in FDA regulation of software as a medical device could be applied to ChatGPT. “As we become more familiar with the technology, there might be changes that are made specifically for this,” Peloquin said. “But it strikes me that a lot of it is going to be following on the principles that already exist.”

Given that FDA already approves medical devices using locked artificial intelligence, the next stage will be approval of devices that incorporate learning artificial intelligence, according to New York University professor of regulatory law and policy Catherine Sharkey. She said the agency clearly has authority to regulate the use of large language models that are incorporated into medical devices or drugs.

“There can be questions like those raised with direct-to-consumer genetic testing, like the boundary line between what’s a medical device and what’s just information that they’re regulating,” Sharkey said. “My own view is that FDA has rather broad authority.”