The FTC has settled a case with a company operating an artificial intelligence-enabled “writing assistant” service. Among the use cases Rytr LLC offered was one that would craft testimonials and reviews, allowing users to select desired tone (e.g., “formal,” “funny,” “convincing”), keywords and phrases, level of creativity (e.g., “optimal,” “high,” “max”), and the number of reviews sought. Based on this user input, Rytr’s AI service would then generate genuine-sounding, detailed reviews quickly and with little user effort. According to the FTC’s complaint, these reviews would almost certainly be false for the users who copy the generated content and published it online. Rytr charged subscription fees for various levels of service, including an unlimited output subscription. Rytr’s users generated tens of thousands of reviews.

The FTC brought this case under a theory of unfairness, alleging that Rytr’s service is likely to cause substantial consumer harm by “pollut[ing] the marketplace with a glut of fake reviews” and has no or de minimis reasonable, legitimate use. Consumers cannot readily avoid Rytr’s reviews since they appear authentic enough to make it impossible for consumers to distinguish a real review from a fake one. The FTC also alleged that Rytr’s AI review service furnished others with the means and instrumentalities to engage in their own deceptive practices. The FTC’s order bans Rytr from selling any service dedicated to – or promoted as – generating consumer reviews or testimonials.

The FTC vote authorizing the staff to issue the complaint and proposed administrative order against Rytr was 3-2, with Commissioners Melissa Holyoak and Andrew Ferguson voting no. In their dissenting statements, the Commissioners objected to the “misapplication” of the FTC’s unfairness authority to the facts presented, including by failing to offer evidence that any AI-generated Rytr reviews actually were posted online. They also took issue with the majority’s assertion that Rytr’s review service provides no countervailing benefits to consumers, for example, the ability of AI to suggest new lines of thought that may never have occurred to a user in the first place. They cautioned that AI is in its nascency, and that the FTC should not “plunge headlong with aggressive regulation” but rather stay its hand.

This case was part of the FTC’s “Operation AI Comply,” a law enforcement sweep aimed at combating AI deception. The FTC indicates that more is to come, and that the agency is examining whether AI and other automated tools are being used for fraud, deception, unfair manipulation, or other harmful purposes, and on the back end, we’re looking at whether automated tools have biased or discriminatory impacts. Watch this space for additional developments.