Algorithm updates
Mirroring, speed-ups and crops are now "detection evasion" in YouTube's Spam Policy
In late May or early June 2026, YouTube rewrote its Spam Policy. Speeding up audio, heavy filters and crops used to dodge detection can now bring strikes.
Updated October 4, 202610 min read9 sources
In short
No. Between May 28 and June 4, 2026 (dates from Internet Archive snapshots, because YouTube did not announce it), YouTube rewrote its Spam Policy. The new text names "Detection evasion": "speeding up audio, heavy filters, cropping" used to bypass abuse detection, with a horizontally mirrored TV episode as the example. The Spam Policy is part of the Community Guidelines, so a breach can bring a warning, strikes or termination, not only lost monetization.
On this page
- YouTube rewrote its Spam Policy in late May or early June 2026, without an announcement
- Old rules and new rules side by side
- What "detection evasion" means, and what it does not mean
- Which common clip edits are now risky?
- Why a Spam Policy breach is worse than lost monetization
- How the Spam Policy connects to the October 2026 Shorts update
- Does campaign permission protect you?
- What to do now: a step-by-step check for clip channels
- What YouTube has not said
- Questions
- Sources
YouTube rewrote its Spam Policy in late May or early June 2026, without an announcement
YouTube's spam rules live on one Help Center page (answer 2801973). The Internet Archive's Wayback Machine stores dated copies of that page. We compared them.
The copy from May 28, 2026 still has the old title, "Spam, deceptive practices, & scams policies", and the old lists. The copy from June 4, 2026 has a new title, "Spam Policy", and a new list of named categories. So YouTube changed the page at some point between those two dates. The text in the June 4 copy is the same as the live page on October 4, 2026.
We found no blog post about the change. YouTube's changelog of policy updates lists other 2026 changes but not this one. The date window therefore comes only from the archive snapshots, not from YouTube. If you read that the rule changed "in June", treat that as an estimate with the same weakness.
Old rules and new rules side by side
The old page already banned scraping other creators at scale. The new page is shorter and more specific about how clip channels behave. These are the parts that matter for clippers:
| Topic | Old page (copy of May 28, 2026) | New page (copy of June 4, 2026, still live) |
|---|---|---|
| Scraped clips | "Massively uploading content that you scraped from other creators" | Scraped content: re-posting other videos "without adding anything of your own". Example: a channel that "takes popular clips and re-uploads them in their entirety" |
| Edits that hide a re-upload | Not named | Detection evasion: "speeding up audio, heavy filters, cropping" when "designed to bypass abuse detection". Example: a TV episode "mirrored horizontally to avoid detection" |
| Volume | "Posting the same content repeatedly across one or more channels" | Automated or synthetic mass-production: using tools or AI to "churn out high volumes of similar content with minimal changes" |
| Clickbait | "Misleading metadata or thumbnails" | Malicious clickbait. Example: a title that promises a full sports match when the video "shows only a clip" |
| Penalties | Monetization suspension, warning, strike, termination | Unchanged: the same warning, strike and termination text |
Two details are new and specific. First, the new page says the policy applies to "coordinated networks of channels", not only to single channels. Second, the mass-production category ends with "technical manipulation to bypass filters or trick viewers". So technical manipulation is named in two categories, not one.
What "detection evasion" means, and what it does not mean
The rule is about purpose. YouTube does not ban cropping, filters or a faster audio track as such. It bans those edits when they are "designed to bypass abuse detection". A crop that turns a 16:9 podcast frame into a 9:16 Short is a normal edit. A zoom, a mirror and a pitch shift added to a clip only so that YouTube's matching systems do not recognise it are the target.
YouTube does not say which systems it means by "abuse detection". This is our reading; YouTube has not said it: the mirrored TV episode example points at matching systems such as Content ID, which compares uploads with reference files from rights holders. A mirror flip or a speed change has no creative purpose in a TV episode. Its only job is to break the match.
YouTube also does not say how it decides what an edit was "designed" to do. It cannot read your mind, so in practice the pattern of the channel is the evidence. A channel where every upload is a full, unchanged moment from another creator, flipped and sped up, tells a clear story. A channel that crops to vertical, adds captions and cuts a 40-minute segment down to its best 45 seconds tells a different one. This is our inference from the policy wording, not a YouTube statement.
Which common clip edits are now risky?
Four techniques are named on the policy page. The risk ratings in the last column are our reading of the purpose test, not YouTube ratings.
| Edit | Named in the Spam Policy? | Risk when its purpose is to hide a re-upload |
|---|---|---|
| Horizontal mirror (flip) | Yes, it is the example | High. It has almost no creative use in a clip of a person talking |
| Speeding up audio | Yes | High. YouTube's monetization rules also list songs changed only in "pitch or speed" as reused content |
| Heavy filters or colour washes | Yes | High when the filter makes the clip worse to watch and adds nothing |
| Cropping or zoom | Yes | Low for a normal vertical reframe. High for an odd crop or zoom that only shifts the frame |
| Borders, frames, blurred background fill | No | Our reading: low on its own; it does not add value either |
| Small speed change of the whole video (for example 1.05x) | Yes, when the audio is sped up too | High. A change viewers do not notice has no creative purpose |
| Re-encoding, noise, changed resolution | No | Our reading: same purpose test; these add nothing a viewer sees |
The safe question is simple: would you make this edit if no detection system existed? If the answer is no, the edit is evidence against you.
Why a Spam Policy breach is worse than lost monetization
Most clip channels worry about the reused content rule. That rule is a monetization policy: if you break it, YouTube can stop paying you, but your videos stay up. The Spam Policy is part of the Community Guidelines, the rules for everyone on YouTube. A breach can remove videos and add penalties to the channel itself.
YouTube's strike system works like this:
| Step | What happens |
|---|---|
| First violation | Usually a warning. An optional policy training lets the warning expire after 90 days; the 90 days start when you finish the training |
| Same policy broken again within that window | The warning may not expire, and the channel may get a strike |
| First strike | 1 week without uploading videos or live streams, creating posts and other actions. The strike stays for 90 days |
| Second strike within 90 days | 2 weeks without posting content. Each strike expires 90 days after it was issued |
| Third strike within 90 days | The channel may be permanently removed |
| Severe abuse | Termination is possible without a warning |
Two more rules matter. Deleting a video does not remove a strike, and YouTube says it may issue a strike on deleted content. And the Spam Policy page says YouTube "may suspend your monetization or terminate your channel or account" for violations, separate from strikes. If YouTube turns off or restricts a channel, its strike page says you may not use, create or acquire another channel to get around the restriction.
How the Spam Policy connects to the October 2026 Shorts update
On October 1, 2026, a TeamYouTube staff member ("JJ") posted in the YouTube Help Community that Shorts recommendations will "further prioritize original content and reduce the reach of content re-uploaded" from other creators. Channels "identified as primarily aggregating or re-uploading" other creators' videos "will likely see less distribution within the Shorts feed." The post says the change is aligned with YouTube's policies, and the word "policies" links to the Spam Policy page.
The same post warns against "minor technical edits or template-based changes simply to alter a video." That is the detection evasion idea in plain words. Read together, the two texts give clip channels three separate risks for the same habit:
- Reach: the recommendation system shows the channel's Shorts less (the October 2026 update). See YouTube's October 2026 originality update.
- Money: reused content is not monetized, and non-original Shorts views do not count toward the Shorts Creator Pool. See YouTube's monetization policies explained.
- The channel itself: warnings, strikes and termination under the Spam Policy.
Earlier signals point the same way. In his annual letter of January 21, 2026, YouTube CEO Neal Mohan wrote that YouTube is building on its spam and clickbait systems to reduce "low quality, repetitive content."
Does campaign permission protect you?
Permission solves a copyright problem. It does not solve a spam problem. The Spam Policy's "Scraped content" category asks whether you added "anything of your own", not whether you had a licence. YouTube's reused content policy says the same in so many words: it "applies even if you have permission from the original creator." See why permission is not a defence on YouTube.
Permission also removes the reason to hide. If a creator really wants clips of their stream on your channel, the clean fix for Content ID claims is not a mirror flip. Rights holders who use YouTube Studio Content Manager can add your channel to an allowlist, and then Content ID "won't claim any videos uploaded to those channels going forward." Existing claims are not released automatically, and each co-owner of an asset must add you separately. Ask the campaign or the creator's team for this.
This is our reading; YouTube has not said it: a clipper who uses evasion edits on a clip they are allowed to post gains nothing and still creates the pattern the policy describes. For copyright itself, see copyright and fair use for clippers.
What to do now: a step-by-step check for clip channels
- Stop the named edits. Remove mirror flips, unnoticeable speed or pitch changes, and heavy filters from your templates. Keep crops only when they reframe the shot for vertical viewing.
- Check your last 30 uploads as a reviewer would. Look at the channel as a whole. If most uploads are full moments from other creators with only technical changes, you match the "Scraped content" example.
- Add something a viewer can see or hear. YouTube names "original commentary, educational value, or unique editing." For clips, that means a selected and tightened moment, context in captions or on screen, and a reason to watch that the source does not give in the same form.
- Do not spread one clip across many channels. The policy now names "coordinated networks of channels" and mass-production "with minimal changes". See running multiple accounts.
- Fix Content ID claims through the rights holder. Ask the campaign for an allowlist entry or a claim release instead of editing around the match.
- Read every warning email. A first warning is a chance to take the policy training and reset. A second violation of the same policy within the window can become a strike.
- Do not delete and re-upload to escape a warning. Deleting does not remove a strike, and re-uploading the same clip is itself a spam pattern.
What YouTube has not said
Several questions are open. We list them so you do not mistake a guess for a rule.
- YouTube has not given the exact date of the rewrite. The May 28 to June 4, 2026 window comes from archive snapshots.
- YouTube has not said whether the new categories brought new automated enforcement, or only new wording for existing rules.
- YouTube has not named the detection systems behind "abuse detection" or published a threshold for crops, filters or speed changes.
- YouTube has not said how it treats old uploads made before the rewrite. The policy page applies to "all types of content on YouTube" and does not mention a cut-off date.
Until YouTube says more, the safest assumption is that the written rule applies to everything on your channel now.
Questions
Is cropping a landscape clip to vertical now against the rules?
No. The Spam Policy targets cropping "designed to bypass abuse detection". Reframing a 16:9 shot for a 9:16 Short is a normal edit with a clear purpose. An odd zoom added only to shift the frame is the risky case.
When exactly did YouTube change the Spam Policy?
YouTube did not announce it. An Internet Archive copy of the page from May 28, 2026 shows the old text, and a copy from June 4, 2026 shows the new text, so the change happened between those dates.
Can a Spam Policy violation get my channel a strike?
Yes. The Spam Policy is part of the Community Guidelines. The first violation is usually a warning; repeat violations can bring strikes, and three strikes within 90 days may end the channel. Severe abuse can lead to termination without a warning.
Is the Spam Policy the same as the reused content policy?
No. Reused content is a monetization policy: breaking it can cost you YouTube earnings. The Spam Policy is a Community Guidelines policy: breaking it can remove videos and add warnings or strikes. A clip channel can break both at once.
Do I need to delete old clips that I flipped or sped up?
YouTube has not said how it treats older uploads. The policy text has no cut-off date. If a channel has many such uploads, removing or replacing them reduces the pattern a reviewer would see. Deleting does not remove a strike already given.
Does a Content ID allowlist make my clips original?
No. An allowlist only stops Content ID claims from that rights holder. It does not change the Spam Policy or the reused content policy, which ask whether you added something of your own.
Sources
We checked every claim on this page against these sources on the date above. Platform rules and rates change often. Found an error? Tell us.
- 1YouTube Help: Spam Policy (live page)support.google.com
- 2Wayback Machine: Spam policy page, copy of May 28, 2026 (old text)web.archive.org
- 3Wayback Machine: Spam Policy page, copy of June 4, 2026 (new text)web.archive.org
- 4YouTube Help Community: Prioritizing original content on Shorts (October 1, 2026)support.google.com
- 5YouTube Help: Community Guidelines strike basicssupport.google.com
- 6YouTube Help: YouTube channel monetization policiessupport.google.com
- 7YouTube Help: Exempt channels from Content ID claims (allowlist)support.google.com
- 8YouTube Help: Updates to YouTube's policies (changelog)support.google.com
- 9YouTube Blog: Neal Mohan, The future of YouTube 2026blog.youtube
Related guides
- Shorts originality updateWill my clip channel's Shorts lose reach after YouTube's October 2026 originality update, and what counts as adding value?
- Permission isn't a defenceThe creator paid me to clip them. Why can my YouTube channel still be demonetized or down-ranked?
- Monetization policiesWhich YouTube monetization policy will a reviewer cite against my clip channel, and what do they look at?
- Copyright and permissionIs clipping legal? Copyright, permission and fair use for clippers