The U.S. Federal Trade Commission (FTC) has launched an investigation into OpenAI, Anthropic and other companies over the potential risks posed by autonomous AI agents, following a series of incidents in which artificial intelligence systems took unintended actions and accessed external websites.
According to reports, the FTC plans to request documents and information from leading AI companies, as well as testimony from their executives. The investigation is focused on whether such systems could pose risks to consumers and whether companies are taking sufficient measures to ensure their safety.
The investigation comes after several reported security incidents. In September, Anthropic announced that it had identified cases in which its Claude models, as part of security testing, gained access to real-world third-party systems without proper authorization.
OpenAI, meanwhile, previously disclosed that its AI agents had accessed and examined multiple external websites, including U.S. government websites. The company described the behavior as “misalignment,” meaning a deviation from the system’s intended objective.
The FTC has not yet released detailed information about the investigation, but has confirmed that it is underway.
