Short answer: no, not if they are presented as coming from real customers. Since October 2024 the US Federal Trade Commission’s rule on reviews and testimonials has banned fake testimonials, and when the FTC announced it, it named AI-generated fake reviews as an example. But “AI” covers a lot of tools, and many of them are perfectly fine to use.
This is a plain summary, not legal advice. The FTC’s questions and answers on the rule are the best primary source.
What the rule actually bans
The rule (16 CFR Part 465) bans reviews and testimonials that misrepresent who gave them or what their experience was. It does not matter how the fake was made. A testimonial typed by an intern, bought from a review farm, or generated by a chatbot breaks the rule in the same way, because it claims an experience that a real person did not have.
It also reaches beyond the business that publishes the testimonial. Agencies and services that write or sell fake testimonials are covered too, and courts can impose civil penalties for knowing violations. Our plain English guide to the rule covers the other five things it bans.
What the FTC says about AI avatars
The FTC has said the rule has no blanket ban on AI-generated avatars in marketing, and that it was written so as not to prohibit virtual influencers. An obviously artificial presenter explaining your product is not a consumer review.
The line is crossed when an avatar is presented as a customer and speaks about an experience nobody had. That is a fake testimonial. The FTC also points out that avatars can be deceptive under the wider FTC Act even when the rule itself does not apply, for example a celebrity likeness used without permission.
Where common AI uses fall
| Use of AI | Risk | Why |
|---|---|---|
| Transcribing a recording | Safe | The words are the customer’s own. |
| Captions from the transcript | Safe | As long as they match what was said. |
| Translating, labelled as a translation | Usually safe | Keep the original too. |
| Choosing which real clips to show | Safe if representative | Showing only cherry picked praise can still mislead. |
| Fixing typos in written text | Low risk | Better still, let the customer approve it. |
| Rewriting or “improving” a testimonial | Risky | The words are no longer theirs. Get their approval or don’t. |
| Summarising many reviews into one quote | Risky | It reads as one person who never spoke. |
| Writing testimonials from scratch | Banned | A fake testimonial, however it was made. |
| An AI avatar presented as a customer | Banned | An invented person claiming an experience they never had. |
The FTC has not published guidance on AI editing of a real customer’s text. The ratings above follow the rule’s core test: does the published testimonial still truthfully show who said it and what they experienced?
The test to apply
Before you publish anything AI has touched, ask three questions:
- Is there a real customer behind it, who actually used what you sell?
- Are these their words, or words they have seen and approved?
- Would they recognise it as what they meant?
If any answer is no, don’t publish it.
How Akteora uses AI, and where it doesn’t
We build testimonial software, so the rule applies to tools like ours as much as to our customers. Akteora uses AI in a few narrow ways and refuses it everywhere else:
- Transcripts of video and audio testimonials, so you can search and caption them.
- Reels: a language model suggests which moments of real recordings to use. It returns timestamps, never words, and each one is checked against the transcript before it is used. Captions are the transcript’s own words.
- Written testimonials: an optional tidy up is shown to the customer, and only used if they approve it before sending. Their original is always kept.
- Never: no generated testimonials, no rewritten recordings, no invented people. Originals are never edited or overwritten.
That way the honest path is also the easy one. Start free and collect testimonials from real customers, or read how to ask for one.
Common questions
Can I use AI to clean up a customer’s written testimonial?
Fixing typos is low risk, but rewriting changes whose words they are. The safest approach is to show the customer the tidied version and use it only if they approve it, keeping their original as well.
Are AI avatars in ads allowed?
The FTC says the rule has no blanket ban on AI-generated avatars. An avatar presented as a real customer with an experience they never had would be a fake testimonial, and avatars can also be deceptive under the wider FTC Act.