The Federal Trade Commission is investigating whether OpenAI’s ChatGPT has harmed people by publishing false information about them, posing a potential legal threat to the popular app that can generate eerily human-like content using artificial intelligence.
In a civil subpoena to the company made public Thursday, the FTC says its investigation of ChatGPT focuses on whether OpenAI has “engaged in unfair or deceptive practices relating to risks of harm to consumers, including reputational harm.”
One question asks the company to “describe in detail the extent to which you have taken steps to address or mitigate risks that your large language model products could generate statements about real individuals that are false, misleading or disparaging.”
The new FTC investigation under Chair Lina Khan marks a significant escalation of the federal government’s role in policing the emerging technology. But it could mean another risky venture into uncharted territory for the agency, which has suffered recent legal setbacks in its antitrust enforcement efforts.
“When ChatGPT says something wrong about somebody and might have caused damage to their reputation, is that a matter for the FTC’s jurisdiction? I don’t think that’s clear at all,” said Adam Kovacevich, founder of Chamber of Progress, an industry trade group.
Such matters “are more in the realm of speech and it becomes speech regulation, which is beyond their authority,” he said.
OpenAI didn’t respond to requests for comment.
The FTC has broad authority to police unfair and deceptive business practices that can harm consumers, as well as unfair competition, but critics say Khan has sometimes pushed its authority too far—as illustrated by a federal judge’s decision this week to dismiss the FTC’s attempt to block Microsoft’s acquisition of Activision.
On Thursday, Khan came under fire for her agency’s investigation of Twitter’s privacy protections for consumers. Republicans say the probe was driven by progressives angry over Elon Musk’s takeover of Twitter and his loosening of content moderation policies. Khan responded that the agency was only interested in protecting the privacy of users.
In its civil subpoena to OpenAI, the FTC asked the company detailed questions about its data-security practices. It cited a 2020 incident in which the company disclosed a bug that allowed users to see information about other users’ chats and some payment-related information.
Other topics covered by the FTC subpoena include the company’s marketing efforts, its practices for training AI models, and its handling of users’ personal information. The letter was reported earlier by The Washington Post.
The Center for Artificial Intelligence and Digital Policy, a think tank, filed a complaint with the FTC in March concerning ChatGPT, terming it “biased, deceptive and a risk to privacy and public safety,” and arguing that it satisfies none of the FTC’s guidelines for AI use.
The Biden administration has begun examining whether checks need to be placed on artificial-intelligence tools such as ChatGPT. In a first step toward potential regulation, the Commerce Department in April put out a formal public request for comment on what it called accountability measures.
Lawmakers in both parties—led by Senate Majority Leader Chuck Schumer (D., N.Y.)—also have made regulating artificial intelligence a priority for the current Congress.
In addition to concerns about potential reputational risks, lawmakers say they also worry that AI tools can be abused to manipulate voters with disinformation, discriminate against minority groups, commit sophisticated financial crimes, displace millions of workers or create other harms. Lawmakers have been especially concerned about the risks of so-called deep fake videos that falsely depict real people taking embarrassing actions or making embarrassing statements.
But new legislation or other measures are likely months away, if not longer. And lawmakers must worry that any significant action they take will risk slowing the pace of U.S. innovation, in what is shaping up as a vital competition with China to dominate the markets for AI tools.
Even ChatGPT’s creators have urged more government oversight of AI development.
In a hearing before Congress in May, OpenAI CEO Sam Altman called on Congress to create licensing and safety standards for advanced artificial-intelligence systems, as lawmakers begin a bipartisan push toward regulating the powerful new tools available to consumers.
“We understand that people are anxious about how it can change the way we live. We are, too,” Sam Altman said of AI technology at the Senate subcommittee hearing. “If this technology goes wrong, it can go quite wrong.”
