FTC Serves Subpoenas on OpenAI, Anthropic Amid Claims of Consumer Safety Failures

On Tuesday, the Federal Trade Commission disclosed that it has served subpoenas on two top AI companies—OpenAI and Anthropic—initiating an investigation into potential consumer damage linked to recent safety lapses.

The commission’s probe centers on whether the firms’ flagship language models have been used in manners that might deceive users, generate harmful material, or breach consumer‑protection laws. The subpoenas seek internal records, communications and testing data concerning the development, evaluation and launch of the companies’ most popular AI offerings.

OpenAI and Anthropic have come under intensified examination following several high‑profile episodes where their models produced prohibited content, offered erroneous medical guidance, or were employed to create convincing misinformation. Although both firms have publicly vowed to strengthen safety measures, regulators contend that the swift market growth of generative AI could outstrip current oversight frameworks.

The commission’s action represents a significant step up from prior, informal AI‑safety inquiries. While the FTC has traditionally exercised its powers to curb deceptive conduct in emerging tech, this constitutes one of the earliest formal subpoenas aimed at major AI developers. Its investigative authority enables it to demand evidence, question staff and determine if consumer‑protection statutes have been violated.

Analysts suggest the subpoenas may reflect expanding regulatory focus on the industry, particularly as AI utilities become integrated into consumer products such as chat assistants and content‑creation tools. The findings of the FTC’s inquiry could shape how firms craft risk‑mitigation approaches, reveal model constraints, and interact with supervisory agencies. OpenAI and Anthropic have both said they will work with the probe while further honing their safety procedures.

Legal experts warn that any determination of non‑compliance might trigger enforcement measures such as penalties or required alterations to product design. Yet the FTC has not provided a schedule for wrapping up the investigation, leaving both firms and their users in a state of uncertainty. This development highlights the mounting clash between swift AI progress and the demand for strong consumer safeguards.