// HOW-TO · SHORT-FORM

How to tell original Shorts content from reposted clips (2026)

Tell if your Shorts count as original content or reposted clips before you publish: a surface-signal audit and the replaceability test YouTube now applies.

Last verified · 2026-10-03 · by Moe Ameen

As of October 1, 2026, whether a Short reads as original or reposted is no longer a philosophical question — it has a reach consequence. YouTube Creator Liaison Rene Ritchie announced on the Creator Insider channel that the Shorts recommendation system is being updated to prioritize original Shorts and reduce the reach of channels that mainly aggregate or re-upload other creators' clips without significant changes. So before you publish, it's worth knowing which side of the line a given Short actually sits on.

This is a diagnostic walkthrough, not a fix-it one. The goal is to look at a finished or planned Short and classify it honestly — original, borderline, or reposted — the way a recommendation system that can't read your intent would. Once you have that verdict you know whether to publish as-is, rework it, or scrap the approach. If the verdict is "needs work," the remediation steps are in [how to repurpose clips without losing YouTube reach](/how-to/repurpose-clips-without-losing-youtube-reach); the full map of the distinction is the guide on [original Shorts content vs reposted clips](/guides/original-shorts-content-vs-reposted-clips).

The steps

Ask the dependency question first. The fastest classifier is one question: if you deleted the borrowed clip, would there be a video left? If the answer is no — the source footage is the whole substance — you're starting at the reposted end no matter how much you dressed it up. If the answer is yes, because your commentary, your footage, or your argument could stand on its own, you're starting at the original end. Everything after this step is refining where between those ends the piece lands.
Separate reach from monetization before you judge. Decide which question you're actually answering, because they use different rules. The October 2026 change is about reach — how widely the recommendation system shows the Short. YouTube's reused-content and inauthentic-content policies are separate monetization rules about whether the channel earns ad revenue. They share the principle that borrowed content needs significant change, but a Short can lose reach without being demonetized and vice versa. This audit judges reach; use [make your content original enough to monetize](/how-to/make-youtube-content-original-for-monetization) for the money side.
Read the surface the way the algorithm does. The recommendation system judges what it can see, so inventory your changes and sort them. Surface-only edits that DON'T count toward originality: captions, borders, filters, a speed change, a voiceover describing what's on screen, and the same intro/caption/cut template stamped across many posts — Ritchie named VO descriptions, minor technical edits, and template-based bulk changes directly. Substance edits that DO count: commentary with an argument, analysis, re-editing that changes what the video is about, and your own footage. If all your changes are in the first bucket, it's reposted.
Run the replaceability test. Ask honestly: could another creator download the same source clips and reproduce this video with a generic treatment? If yes, it's replaceable, and replaceable reads as reposted — any ten channels could have made it. If no, because the take, the edit, the framing, or the footage is specifically yours, it's transformed. "Replaceability" is a creator-community framing, not YouTube's official term, but it matches the surface-judgment logic closely enough to use as a hard gate.
Score the substance ratio. Estimate what share of the finished video is your substance versus borrowed substance — not screen time of the clip, but where the value lives. A reaction that is 90% someone else's clip and 10% "wow" is heavily borrowed. A breakdown where the clip plays under your explanation and your explanation is the reason to watch is heavily yours. The closer the ratio tips toward borrowed, the closer the Short sits to the reposted end that loses reach. There's no official threshold — the point is to force an honest read, not hit a number.
Check the channel pattern, not just this one Short. The announcement targeted channels that MAINLY aggregate or repost, so classify your feed, not only the video in front of you. One well-transformed reaction on a channel full of your own original work is a different signal than the same reaction on a channel that is wall-to-wall reposts. If most of your library is barely-changed clips, even your genuinely original pieces are swimming against the channel-level pattern — and Ritchie noted a channel shifting toward original work gets re-evaluated, though no timeline was given.
Turn the verdict into an action. Resolve each Short to one of three outcomes. Original (survives the dependency question, substance is mostly yours, not replaceable): publish. Borderline (depends heavily on the clip but has a real seed of your own): rework it — add depth, re-edit the core, bring your own footage — before publishing. Reposted (pure re-skin): either transform it properly or drop it, because publishing it as-is spends an upload slot on something built to be under-recommended.

Common gotchas

Counting effort instead of addition. You can spend an hour on captions, a border, and a speed ramp and still score as reposted — the audit measures what you added, not how long the editing took.
Grading the clip's screen time instead of where the value lives. A short source clip under a thin voiceover is still reposted; a long clip under a carrying argument can be original. Judge substance, not duration.
Treating a verdict as permanent. A borderline Short isn't a failure — it's a rework candidate. Most reposted-leaning pieces can be pushed to the original end with a real commentary or re-edit layer.
Auditing one Short and ignoring the channel. The change reads the channel pattern; a single original video can't offset a feed built mainly on reposts.
Confusing this with a monetization check. A clean reach verdict doesn't mean the channel clears the reused-content monetization policy — they're separate systems with separate consequences.
Assuming AI involvement decides it. An AI-made Short can be either original or reposted; the verdict turns on whether your point of view or material is in it, not on whether a model was used.

Where Kompozy fits

This audit is really a classification step that most creators do in their head, inconsistently, right before hitting publish. Kompozy is a full AI content generation and multi-platform publishing engine, and the useful thing it does for this workflow is make that classification a standing checkpoint instead of a last-second gut call. Every piece the engine generates waits in a per-post review queue for your sign-off before it ships — which is exactly the moment to run step seven's verdict. Nothing fans out until you've decided it's on the original side, and [Autopilot](/glossary/autopilot) is opt-in per source only for the ones you've already judged safe to run hands-off.

The deeper fit is upstream of the audit. The formats that pass the dependency question in step one by construction are the net-new ones with no borrowed source at all — a [Quote Graphic, Carousel, blog post, or newsletter](/glossary/output-buckets), a [Persona Tweet](/glossary/persona-tweet), or an avatar scene video — all generated from your [Persona Brief](/glossary/persona-brief), which holds your point of view and phrasing so the substance is yours from the first draft. "Would this survive without a source clip?" answers itself when the clip was never the substance. When you do build on outside footage, [Clipped Shorts](/glossary/clipped-short) re-cut and re-hook it and [Persona Shorts](/glossary/persona-shorts) layer your actual commentary on top, which is the kind of addition the audit scores as original rather than the re-skin it flags as reposted.

Honest boundary: Kompozy can't pass the audit for you. The point of view in step three, the firsthand material in step five, and the final verdict in step seven are human judgment this tool deliberately keeps in your hands. What it removes is the cost of producing on the original side every time, and the risk of one platform owning your distribution — the same approved piece publishes across the eight social platforms plus blog and email, so a reach re-evaluation on Shorts never takes everything with it. Starter is $199/mo (5,500 credits) for a solo creator; Pro is $499/mo (18,000 credits) for daily, multi-platform output with autopilot; Enterprise is custom.

Frequently asked questions

How do I know if my Short counts as reposted?

Run the dependency question: if you removed the borrowed clip, is there a video left? If not, it's reposted. Then check whether your only changes are surface edits — captions, borders, filters, a speed change, or a voiceover of what's on screen — which YouTube says don't count. If the substance is someone else's and your changes are cosmetic, the recommendation system reads it as reposted and reduces its reach.

What is the replaceability test?

A one-question gate creators use since the October 2026 change: could another creator download the same source clips and reproduce your video with a generic treatment? If yes, it's replaceable and reads as reposted. If no — because the commentary, edit, framing, or footage is specifically yours — it reads as original. It's a community framing of YouTube's enforcement logic, not an official term, but it tracks the surface-judgment well.

Does a reposted verdict mean I lose monetization?

Not necessarily. The October 2026 reach change and YouTube's reused-content monetization policy are separate systems. A Short can be under-recommended for reach while the channel stays monetized, and a channel can pass reach but fail a monetization review. They share the principle that borrowed content needs significant original change, but one governs how far a Short travels and the other governs whether the channel earns ad revenue.

Can an AI-generated Short pass as original?

Yes, if it carries your point of view or your material. Where an AI Short lands on the line depends on what it contains, not on the fact that AI made it. A clip re-skinned with an AI voiceover is reposted; an AI-narrated short delivering your actual argument, or a net-new format generated from your own brief, is original. The model is a production tool — the originality still has to come from you.

When did the originality reach change take effect?

YouTube didn't publish a start date. The Creator Insider video announcing it on October 1, 2026 didn't give a date for the update or say how long a channel's re-evaluation takes after it shifts toward original content. Treat the change as in effect and ongoing rather than waiting for a specific switch-on date, and audit your Shorts now.