On 27 March 2025 the US Federal Trade Commission filed suit against Cleo AI, Inc., alleging that it falsely advertised a cash advance service that could be provided quickly. According to the FTC, its advertising promised funds “today” or “instantly”, but some consumers received far less than the advertised amount, and there were hidden fees and a subscription mechanism that was difficult to cancel.
An AI app is not a regulatory exemption
The first signal from the case is that using AI technology in a product does not change the application of advertising law or consumer protection law. The FTC’s allegations concern the truthfulness of claims, fee disclosure and subscription cancellation, all of which fall within traditional regulatory territory and have nothing to do with the algorithm itself.
For companies going overseas, this means that in marketing copy and subscription flows, being new in technology cannot be a reason to relax scrutiny.
The gap between the claim and the reality
The core of the FTC’s concern is the gap between the claim and the reality: the promised arrival time did not match the actual one, and the advertised amount did not match the actual one. Problems of this kind are especially easy to produce in AI products, because generative capability makes marketing copy cheap to produce while the review step has not been strengthened in step.
See our note on Korea’s punitive damages regime for false AI advertising; other jurisdictions are also building stricter liability mechanisms for AI-related misleading advertising.
Compliance points in subscription design
A subscription that is difficult to cancel has long been a focus of FTC enforcement. Common risky designs include a cancellation entry point buried too deep, multiple confirmation steps, continued charges after cancellation, and automatic renewal induced by a free trial.
See our notes on subscriptions and the token economy: in membership, points and subscription products, the design of the cancellation path should be listed separately as a compliance check item.
Three reminders for AI application companies
First, keep a review trail for marketing copy. AI-generated copy needs human review, and the review record should be kept. See the AI Plus deployment in consumer applications.
Second, fees and arrival times must be verifiable. A specific promise in marketing should be backed by corresponding system capability. AI products with money-related functions also need to look at the guidance on AI security in the financial sector.
Third, the cancellation path should be at least as easy as the sign-up path. This is a shared requirement across most jurisdictions and a high-frequency inspection point in enforcement.
The compliance risk of an AI application often lies not in the algorithm itself but in the details of product design and marketing expression.
This article is based on the US Federal Trade Commission’s enforcement action against Cleo AI, Inc. and public information and is for reference only; it does not constitute legal advice.