AI Shorts lost views after Oct 2, 2026? Diagnose first
Not every drop is the originality update. A diagnosis order for AI Shorts: copied clips, sameness, labels, then ordinary causes, using only what YouTube says.

YouTube said on October 2, 2026 that it is updating its recommendation systems to reduce the reach of content re-uploaded from other creators without adding anything of your own (9to5Google, read 2026-10-05). If your views dropped around that date, the update is one possible cause among several. Change things in the right order.
Step 1: are any Shorts built on other people's clips?
This is the named target. List every Short that includes footage from another creator. If the clip is only cropped, filtered or sped up, YouTube's advice, as reported by Android Headlines, is to add commentary, unique edits or storytelling rather than minor technical edits or template-based changes. If you cannot add that, retire the Short.
Step 2: do your Shorts look alike?
Put your last 20 thumbnails and first lines side by side. YouTube's monetization page describes generic or unoriginal templates. A recommendation system is not the monetization system, and nothing we read says it uses the same test, but sameness is a poor sign either way. Check openers, length and structure.
Step 3: labels
On May 27, 2026 YouTube moved the disclosure label for photorealistic AI content to an overlay on Shorts, and said it will apply the label itself when its systems detect significant photorealistic AI use that was not disclosed. A label is not a penalty, and YouTube has not said it lowers reach. If you meant to label a Short and did not, fix it, since repeated non-disclosure can lead to penalties.
Step 4: the ordinary causes
- Topic fatigue: your niche may be saturated this month.
- Retention: did average view duration fall on the affected Shorts?
- Upload cadence: did you change schedule or length mix?
- Seasonality: compare the same week last year if you have it.
What not to do
Do not delete your back catalog in a panic, and do not trust any post that gives a precise rule for what the update penalizes. YouTube has not drawn a line, and 9to5Google says so. Make one change at a time, note the date, and compare a week later.
Write the finding down
Whichever step turns up a cause, note the Short, the cause and the change. After a week, you will know whether the change helped, and you will not be guessing again the next time YouTube announces something. A short log beats a long theory.
Sources
- 9to5Google: YouTube won't recommend unoriginal Shorts, news event only (read 2026-10-05)
- Android Headlines: YouTube cuts reach of re-uploaded Shorts, news event only (read 2026-10-05)
- YouTube Help: Inauthentic content in monetization policies (read 2026-10-05)
- YouTube Blog: Improving AI labels for viewers and creators (read 2026-10-05)
- YouTube Help: Disclose altered or synthetic content (read 2026-10-05)
Related posts
More in Use cases
- Why TTS plus lip-sync avatars feel fake: Griffin's 2.4% vs 48%
Tavus says its older avatar stack passed as human 2.4% of the time and Griffin-Lite 48%. Here is what that means for TTS-then-lip-sync clips.
- Win-back video for lapsed customers: 30-second avatar script, cost
A win-back email clip made with a Sume avatar: what to say in 30 seconds, what not to promise, and the cost per lapsed account on standard, plus and max.
- X carousel ads: 2 to 6 media assets, rendered as one Sume batch
X carousel cards hold 2 to 6 media assets. Make the whole set at one aspect ratio in a single Sume Image API call with n, then check each file.
- X image ad specs: 1:1 or 1.91:1, 800px wide, 3 MB, from Sume
X website image cards take 1:1 or 1.91:1, at least 800px wide and 3 MB. Render 2:1 on Sume, crop to 1.91:1 with Pillow, and check size before upload.
Written by Sume