Platform rules

The AI label is not what costs you the money. Templating is.

Most creators worry about the wrong rule. Disclosing AI is a transparency label with no documented reach penalty. The policy that actually removes channels from monetization is the one about mass-produced, templated content, and it applies whether a human or a model made it.

Policy summary checked on 12 August 2026

YouTube: when you must disclose, and when you must not bother

YouTube's requirement is about realism, not about tooling. In its own words, creators must disclose content that “makes a real person appear to say or do something they didn't do,” “alters footage of a real event or place,” or “generates a realistic scene that didn't actually occur.” It also states that creators “don't need to disclose non-realistic content that's made with AI, or edits to realistic content that are minor.”

Disclose

A real person appears to say or do something they did not do
Footage of a real event or a real place has been altered
A realistic scene is generated that never actually happened

No disclosure needed

Content that is clearly not realistic, including animation and character work
Beauty filters and colour adjustment
Generated captions and subtitles
Script writing, idea generation, and other production assistance

A talking fruit episode sits firmly in the second column: it is animated character work that nobody could mistake for footage of a real event. Where a real product, a real place, or a real person enters the frame, the calculus changes and the honest move is to label it. YouTube also warns that creators who consistently do not disclose “may be subject to manual application of a label, or penalties from YouTube, including removal of content or suspension from the YouTube Partner Program.”

The policy that actually ends channels

In July 2025 YouTube renamed its repetitious content policy to inauthentic content. The target is mass-produced and templated output: videos built from the same formula with little variation, replicable at scale, with no real authorship in them. YouTube's own framing is that this content was never eligible for monetization; the rename made the enforcement explicit.

This is the rule that catches AI channels, and it catches them for a reason that has nothing to do with AI. A channel that posts forty videos with the same voice, the same structure, and the same stock visuals fails it. So does a human doing the same thing by hand.

What passes, in practice

A recognisable cast that returns, so episodes are chapters rather than clones
One real idea per episode: a conflict, a reveal, a joke that only works once
Variation you can see in the first second, not just in the title
Your own packaging: titles, thumbnails, and descriptions written per episode
A disclosure label wherever a real person, place, or product appears

TikTok and Instagram

TikTok's rule is disclosure, not prohibition. Its community guidelines require synthetic media showing realistic scenes or people to be labelled, and it applies AI labels automatically through Content Credentials and its own detection as well as through the creator-applied toggle. Instagram and Facebook operate the same way through Meta's AI Info labelling.

Where the public sources disagree

Whether AI-assisted content qualifies for TikTok's creator payout programmes is reported inconsistently across secondary sources, and the programme terms have changed more than once. Do not take a blog's word for it, including this one: read the current programme terms in the app before you build a business on that revenue line.

Primary sources

Policies change. These are the pages to check rather than a summary, including this one. Nothing here is legal advice.

Questions creators ask before they start

Can you monetize AI-generated videos on YouTube?

Yes. YouTube does not ban AI-generated content from the Partner Program. What it bans is inauthentic content: mass-produced or repetitious videos made from a template with little variation between them. An AI video with a real idea behind it is monetizable; a hundred near-identical AI videos are not, whether a human or a model made them.

Do I have to label a talking fruit video as AI?

Usually not on YouTube. The disclosure requirement covers realistic synthetic content: a real person appearing to say something they did not say, altered footage of a real event or place, or a realistic scene that never happened. An obviously animated fruit character is none of those. TikTok's labelling rule is also aimed at realistic synthetic media. Label anything that could be mistaken for real footage, and when in doubt, label it: the label is a transparency signal, not a penalty.

What actually gets an AI channel demonetized?

Volume without variation. YouTube renamed its repetitious content policy to inauthentic content in July 2025 and enforcement targets templated output at scale. In practice that means the same script structure, the same voice, the same visual formula, published dozens of times. The defence is a recognisable series with characters, an idea per episode, and real editorial choices, not fewer uploads.

Does labelling content as AI reduce its reach?

The platforms describe the label as a transparency notice for viewers rather than a ranking signal. Neither YouTube nor TikTok publishes a documented reach penalty for correctly labelled content. Treat any claim of a secret label penalty as unverified.

Do I own the videos I generate?

With Frutti, yes: you own the output you generate and paid plans cover commercial use. Copyright in purely AI-generated material is unsettled in several countries, and it is your responsibility not to imitate a real person or a brand you do not represent. That is a rule about what you generate, not about which tool generated it.

Build a series, not a template farm.

Frutti is built around a recurring cast and one idea per episode, which is the shape the inauthentic content policy is asking for. Write one free and see.

Write an episode free

At a glance

What Frutti is, in facts you can check.

Frutti is an AI video generator that turns one sentence into a complete vertical episode: script, recurring characters, voices, dialogue and captions. The next episode starts from the same cast, so the output is a series rather than a one-off clip.

What it makes: One finished vertical episode from one written sentence: the script, an original cast that returns episode after episode, performed dialogue with distinct AI voices, and burned-in captions.
What you download: A 9:16 vertical video with the captions already burned in, ready to post without an editor.
Episode length: Episodes are built from long dialogue clips, each one continuous shot: Quick 2 clips of 10 seconds, Standard 3 of 10, Long 3 of 15, Epic 4 of 15. Longer runs come from a series of episodes.
Price: Starter $24/month or $7/month billed yearly, Creator $59/month or $18/month billed yearly. Yearly billing is 70% off and is charged once for twelve months. Prices in US dollars.
What is free: Writing a full episode (script, cast and scene list) on the home page and on every format page, with no account and no card. Rendering runs on credits. There is no free render tier and no watermarked tier.
What an episode costs: A Standard episode is 3 continuous clips of 10 seconds, about 30 seconds finished, with the script, the cast, the voices and the captions included, and costs 1200 credits. A Long episode is 3 clips of 15 seconds at higher resolution and costs 1,800. Starter is $24 a month for 1,200 credits: 1 finished Standard episodes, about $24 an episode, or about $7 an episode billed yearly. The same balance is 1 Quick episodes or about 5 single clips. Writing and previewing a script spends no credits.
Continuing a series: A finished episode keeps its cast. The next one opens with the same characters, the same names and the same outfits already written in, in the same format, so a channel keeps one cast instead of inventing a new one every time.
Watermark: None. No export carries a watermark, on any plan, and there is no watermarked free tier to upgrade out of.
Ownership: You own the output you generate. Paid plans cover commercial use, including client work, brand channels and paid advertising.
Single clips: Single clips render on Frutti E1, our own engine, on every plan. A clip costs what its length costs.
Formats: 12 dedicated formats, one generator page each: Talking fruit drama, Italian brainrot, Talking bugs, Talking teeth, AI baby podcast, Talking dogs, Plush vlogs, News desk parodies, Courtroom shows, Horror shorts, Animal drama, Product ads.
Cancelling: Cancel from Settings at any time. Cancelling stops the next renewal and the plan stays active until the end of the period already paid for. Unused credits are not confiscated when a plan is cancelled, expires or renews.
Refunds: A run that fails before producing output refunds its credits automatically. Unused credits stay on the account when a plan is cancelled, expires or renews; nothing is confiscated. Billing errors such as a duplicate charge are corrected when reported within 14 days. Credits spent on a finished render are not returned, and cancelling does not refund the rest of a paid period: the plan stays active until that period ends. Apple purchases follow Apple's refund process.
Who runs it: Fast AI Labs, the operator of frutti.ai. Web billing runs through Stripe, iOS billing through Apple. Support: [email protected].
Languages: English plus 16 other languages.

Best for

Building a channel around a cast that returns every episode, where the writing carries the video.
Faceless short-form volume that still reads as written rather than templated: a full episode from one line, several a week.
Brands that want a product placed inside a story, in the format their audience already watches.

Not for

Art-directing one shot. Frutti decides the shot, the pacing and the casting for you and returns a whole episode; it does not expose a model picker, camera controls or frame-by-frame iteration on a single clip. If the deliverable is one hero shot you want to tune by hand, that is a different kind of tool.
Editing footage you already have. Frutti generates; it does not cut, caption or reframe your recordings. A clipper such as OpusClip, Klap or Vizard does.
Free rendering. The free stage is the writing: script, cast and scene list, no account and no card. Rendering is paid because generation costs real money per second, and there is no watermarked free tier. If the video you want is text-to-speech over gameplay footage, Frutti does not make that format at all.

Verified 12 September 2026. Prices, credits and model counts are read from the configuration that charges for them, not typed in. Policies: refunds and terms. Machine-readable: /api/public/plans and /llms.txt.