FTC investigating ChatGPT-maker OpenAI for possible consumer harm

The Federal Trade Commission is investigating ChatGPT-maker OpenAI to recognize if the institution has violated user extortion laws.

The Washington Post, which archetypal reported the news, published the FTC's 20-page civilian investigative demand (CID), akin to a subpoena outlining cardinal focuses of the probe. A root acquainted with the substance confirmed the authenticity of the papers to CNBC. The FTC declined to comment.

The FTC says successful the papers that the probe volition absorption connected whether OpenAI has "engaged successful unfair oregon deceptive privateness oregon information information practices" oregon "engaged successful unfair oregon deceptive practices relating to risks of harm to consumers, including reputational harm, successful usurpation of Section 5 of the FTC Act."

AI has go a blistery contented successful Washington, with lawmakers trying to recognize whether caller laws are needed to support intelligence spot and user information successful the property of generative AI, which requires monolithic datasets to learn. The FTC and different agencies person emphasized that they already person ineligible authority to prosecute harm created by AI.

The probe is besides an illustration of the FTC being proactive successful its oversight of a comparatively nascent technology, successful enactment with Chair Lina Khan's stated goal of being "forward-looking" and paying attraction to "next-generation technologies."

The CID asks OpenAI to database the 3rd parties that person entree to its Large Language Models (LLMs), their apical 10 customers oregon licensors, explicate however they clasp and usage user information, however they get accusation to bid their LLMs and more. The papers besides asks however OpenAI assesses hazard successful LLMs and however it monitors and deals with misleading oregon disparaging statements astir people.

The CID asks OpenAI to supply accusation astir a bug the institution disclosed successful March 2020 that "allowed immoderate users to spot titles from different progressive user's chat history" and "may person caused the unintentional visibility of payment-related accusation of 1.2% of the ChatGPT Plus subscribers who were active during a circumstantial nine-hour window."

OpenAI CEO Sam Altman has mostly received a lukewarm invited successful Washington up until this point, with lawmakers praising his openness to discussing the exertion and asking for regulations astir it. But immoderate AI experts person warned that policymakers should besides beryllium alert that the institution has its ain incentives successful articulating its imaginativeness of regularisation and urged them to prosecute a divers acceptable of voices.

OpenAI did not instantly respond to a petition for comment.

Subscribe to CNBC connected YouTube.

WATCH: Investing successful AI: What to Consider