FTC fake followers section: what AI content teams must not buy

FTC rule section 465.8 bars buying or selling fake indicators of social media influence. What it means for AI content teams.

4 min readSume
All posts

Section 465.8 of the FTC rule on fake reviews and testimonials, titled "Misuse of fake indicators of social media influence", makes it a violation to purchase or procure fake indicators that materially misrepresent influence for a commercial purpose, when you knew or should have known they were fake. Section 465.7 covers review suppression. An AI content team should read 465.8 as a purchasing rule: generating videos and images is one activity, and buying fake followers or other fake engagement signals is a different one that the rule addresses.

The two sections in plain terms

The rule text at Cornell LII was read on 2026-10-03. This post paraphrases them for orientation and is not legal advice; read the rule text and ask counsel about your specific practice.

FTC rule sections named in this post (read 2026-10-03)
SectionCovers
465.7Review suppression
465.8Misuse of fake indicators of social media influence (buying or selling them)

Where generation ends

A generation API produces media: images, videos, audio and avatar clips. It returns artifacts and a job record. It does not create an audience, and Sume's job results contain the generated files and receipts, not follower or engagement counts.

That separation is the point. Generating a product video with an avatar is content production. Paying a vendor to add followers, likes or views to the account that posts it is a different transaction, and it is the kind of indicator section 465.8 is about.

  • Generate the content; earn the audience.
  • Do not buy followers, likes or views to make a post look popular.
  • Do not pay for or script reviews that did not come from real experience.
  • Keep brand accounts' growth sources auditable.

Keep a record of how content was made

If a regulator or a client asks how an asset came to exist, a record helps. For generated media the useful pieces are the job id, the request id, the model and the output artifact, all of which come back with the job result. Sume's docs say to include the request id when reporting an issue and to keep the job id as your idempotency key for webhook handling.

Pair that with a purchasing policy that lists what the team may buy: generation credits, stock, voice, captions. Engagement metrics are not on the list.

Sources

Related posts

More in Use cases

All Use cases posts

Written by Sume