If you thought your YouTube Shorts feed was already chaotic, 2026 is about to turn the dial way up. YouTube is working on a feature that will let creators generate Shorts using their own AI likeness — essentially, an officially sanctioned “AI twin” that can keep talking, performing, or selling even when the human behind the channel is offline.

YouTube CEO Neal Mohan teased the feature in his 2026 annual letter, saying that later this year, creators will be able to create Shorts “using their own likeness,” without yet explaining what that actually looks like in practice. Will it be a full-on digital clone that can be animated via text prompts, or more like a templated avatar built from a few training clips? YouTube isn’t saying yet, only promising that details and a launch date are “coming soon.”

What we do know is that this isn’t coming out of nowhere. For the past couple of years, YouTube has been quietly turning Shorts into a playground for generative AI. Dream Screen already lets you type something like “a neon cyberpunk street in the rain” and drop that as an AI-generated background behind your selfie video. Then Google’s DeepMind team plugged in Veo 2, its high-end video model, so creators could generate entire short clips from text prompts and slot them directly into Shorts, without shooting anything at all. The “AI likeness” move feels like the next logical step: instead of just generating the world around you, YouTube is starting to generate you.

On paper, this is creator wish-fulfillment. Imagine a daily Shorts schedule where your AI twin records language-specific versions of your explainer for different markets, automatically dubs itself into Spanish and Hindi, and tests five thumbnail hooks — all while you’re asleep. Mohan’s pitch is very much in that direction: AI as a force multiplier, something that makes it easier for one person to behave like a full-blown studio, cranking out more content in more variations without burning out. He’s already framed 2026 as a moment where “creativity and technology are merging” and compared AI tools to earlier shifts like synthesizers in music or Photoshop in photography.

But this all lands in a YouTube ecosystem that’s already wrestling with “AI slop” — the flood of low-effort, auto-generated videos that clog recommendations and make the platform feel spammy. YouTube has shut down channels pushing obviously fake AI movie trailers and says it’s actively tweaking its systems to demote low-quality, repetitive content. At the same time, the company is telling creators not to think small: Shorts are pulling in around 200 billion views a day, and YouTube clearly wants as much of that as possible to be built using its own AI stack.

That tension — between AI as a creativity booster and AI as a content swamp generator — is why YouTube is pairing these new tools with stricter rules around transparency. Creators are now required to disclose when realistic-looking content is “meaningfully altered or synthetically generated,” including when generative AI is used to depict a real person, place, or event. For sensitive topics like news, politics, health, or finance, YouTube even slaps a more prominent label directly on the video. If you don’t disclose and the content risks misleading people, YouTube reserves the right to add a label itself and, over time, potentially penalize repeat offenders.

That policy becomes really important when we’re talking about AI clones of real people. The closer these AI likenesses get to photorealistic, the more they brush up against deepfake territory — even if they’re fully opt‑in and controlled by the creator. YouTube says it’s updating its privacy tools so people can request the removal of AI-generated or other synthetic content that mimics their face or voice in a convincing way. The company is trying to carve out a middle ground: give creators powerful synthetic media tools, but keep some brakes in place so audiences aren’t constantly tricked or impersonated.

We also shouldn’t ignore the economics angle. Shorts are a key part of YouTube’s strategy to compete with TikTok and Reels, and AI is a way to keep that feed endlessly stocked. If your AI likeness can film a dozen versions of a brand integration, or localize a sponsored Short for five countries without booking a single extra shoot day, that’s a big deal for both creators and advertisers. Mohan has already hinted at tools that make it easier to swap or insert brand deals into existing content, and an AI clone fits neatly into that world — the “you” endorsing a product might not even be a take you personally recorded.

Of course, there’s a risk that this all makes YouTube feel less human. Part of Shorts’ appeal is the sense that you’re watching a real person in real time, reacting to trends, telling stories from their messy bedroom or studio. An AI likeness that churns out perfectly on‑brand content might be efficient, but it could also flatten the quirks and imperfections that make creators feel relatable in the first place. If every creator can spin up an AI twin that never gets tired and always stays on message, feeds could start to look like a never-ending wall of polished, synthetic performances.

There are also deeper questions of creative identity. If a viral Short was written by a human, performed by their AI clone, lightly edited by the platform, and then auto‑dubbed into multiple languages, who’s really the author? Creators will have to decide how much of themselves they’re comfortable handing over to automation — and how clearly they want to signal, “This is AI‑me, not real‑me,” especially for younger audiences who may not draw that distinction as sharply. YouTube, for its part, is leaning on labels and updated parental controls as its main answer.

Still, if you zoom out, the direction is obvious. YouTube wants to be the place where you don’t just upload videos, you generate them — clips, backgrounds, even the on‑screen “you” — all inside its ecosystem. AI likeness Shorts are just one more step toward that future. For some creators, that’s going to be incredibly freeing; for others, it’ll feel like crossing a line. But whether you’re excited or uneasy, you should probably get used to the idea that the face talking to you in your Shorts feed may not always belong to a person who actually hit record that day.