FTC reportedly investigating OpenAI, Anthropic over potential consumer risks

The U.S. Federal Trade Commission is reportedly investigating OpenAI Group PBC, Anthropic PBC and other artificial intelligence providers. 

The New York Times today cited a source as saying that the probe covers multiple areas. According to the report, the FTC plans to review whether the companies have engaged in unfair or deceptive practices. Additionally, it’s believed the agency will “probably” investigate whether rogue AI agents have harmed consumers.

Last week, OpenAI revealed that rogue AI agents had posted ChaGPT users’ images to third-party websites on at least 53 occasions. The agents extracted the photos from the company’s AI training dataset. According to OpenAI, the dataset contained user photos because it incorporates prompts sent to the consumer editions of ChatGPT. The chatbot service offers a setting that enables consumers to opt out of AI training.

It’s possible the FTC is also concerned about the manner in which OpenAI processes healthcare data. In July, the company launched a set of features that enables ChatGPT to answer questions about users’ medical records. Healthcare data is subject to stricter regulations than many other types of information.

Earlier this month, the Wall Street Journal reported that the FTC was planning to investigate the impact of ChatGPT and other chatbots on children’s mental health.

It’s possible AI labs’ recent disclosures of AI-driven cyberattacks will also factor into the probe. Rogue AI agents tied to OpenAI and Anthropic have breached multiple organizations’ networks over the past few months. Last week, the former company disclosed the first such incident to have involved a government agency. Rogue OpenAI agents downloaded nonpublic data from Australia’s healthcare statistics agency.

The FTC is reportedly drafting civil investigative demands, or CIDs, as part of the probe. Those are formal requests for information about a potential regulatory violation. The FTC will reportedly issue the CIDs to OpenAI and Anthropic in the coming weeks.

The agency is also expected to scrutinize METR, a nonprofit AI safety lab. The organization helped OpenAI investigate its AI agents’ cyberattack against Hugging Face Inc. earlier this year. More recently, Anthropic hired METR to review cybersecurity incidents involving its own agents.

The FTC is reportedly growing its technology-focused regulatory team to support the probe. Further down the line, the agency is expected to seek testimony from top executives at the companies that it’s securitizing.

FTC investigations can lead to a formal lawsuit. The agency usually asks the presiding judge to issue fines or order the defendants to change their business practices. In many cases, companies implement the changes sought by the agency before going to court.

Photo: John Taylor/Flickr