DooDooLamb News
You can now remix other people’s YouTube Shorts with AI
Brief published May 22, 2026 · Original source published May 20, 2026
Original reporting by Terrence O’Brien at theverge.com.
Automated brief. Verify important details at the original source.
Google's New Feature Lets You Deepfake Yourself Into TikToks, What Could Go Wrong
Google just announced that YouTube Shorts now comes with an AI remix feature that lets you insert yourself into other people's videos using their Gemini Omni model. Because apparently what the internet was really missing was an easier way to violate consent and copyright simultaneously while looking like a complete amateur doing it.
The feature, cheerfully called "reimagine," appears as an option when you hit the remix button on any YouTube Short. Users can now restyle clips to look like anime, slap giant heads onto people, or digitally insert themselves into content they had absolutely nothing to do with creating. It's like being that person who Photoshops themselves into vacation photos they weren't invited on, except now Google is handing you professional-grade tools to do it at scale.
Google's Gemini Omni is the company's multimodal AI model that can understand and generate across text, images, audio, and video. Think of it as ChatGPT's overachieving cousin who went to film school and won't shut up about it. The model can analyze what's happening in a video and then generate new content that theoretically matches the style, lighting, and context of the original. In practice, this means you can take someone's carefully crafted dance video and turn it into what looks like a rejected Disney Channel Original Movie from 2003.
The technology behind this is genuinely impressive from a computer science perspective. Training an AI to understand spatial relationships, lighting conditions, and visual consistency across video frames is legitimately hard. Google's engineers have essentially created a system that can watch a video, understand its visual grammar, and then generate new content that (mostly) follows the same rules. It's the kind of breakthrough that would have been science fiction five years ago and is now being deployed so teenagers can put their faces on other people's workout videos.
But here's where things get wonderfully dystopian. The feature launches at a time when deepfakes are already causing havoc across the internet, when content creators are fighting tooth and nail for attribution and fair compensation, and when the line between authentic and synthetic content has become so blurred that we need browser extensions to tell us what's real. Google's solution? Let's make it easier for everyone to blur that line even further, but with cartoon filters.
The timing is particularly rich considering YouTube's ongoing struggles with creator compensation and content theft. The platform has spent years trying to build systems that properly credit original creators when their content gets remixed or sampled. Now they're introducing a feature that essentially automates the remixing process while making the original source material even more difficult to identify. It's like spending decades perfecting a system to prevent art forgery and then handing out professional-grade photocopiers at the museum gift shop.
Content creators are already dealing with AI systems scraping their work for training data without consent or compensation. Now they get to watch users actively remix their content using AI tools built by the same platform that hosts their videos. The cognitive dissonance is almost beautiful in its completeness. YouTube has managed to create a system where creators can simultaneously be the product, the customer, and the victim of the same transaction.
The feature also raises fascinating questions about what constitutes fair use in the age of generative AI. If someone takes your video and uses AI to completely transform the style while keeping the basic composition, is that transformative enough to avoid copyright issues? Legal scholars are going to have field days with this one, assuming they can stop laughing long enough to file the briefs.
From a technical standpoint, this represents a genuine advancement in real-time video generation and style transfer. The fact that Google can deploy this kind of capability directly in a web browser without melting users' phones is legitimately impressive. But impressive technology deployed without considering its broader implications is how we ended up with most of our current digital nightmares.
For developers and founders, this feature represents both an opportunity and a warning. The opportunity is clear: if Google can build this into YouTube, similar capabilities will soon be available as APIs for anyone to integrate. The warning is equally clear: the race to deploy AI features is moving so fast that ethical considerations are being treated as optimization problems to solve later.
The reality for most users is that this feature will produce about ninety percent garbage and ten percent content so cursedly good that it goes viral despite nobody being able to explain why. It's the same ratio that governs most user-generated content, except now with the added complexity of not knowing who actually generated what. YouTube has essentially weaponized the uncanny valley and handed it to the same people who brought you the cinnamon challenge.
In the end, Google has created the perfect metaphor for the current state of AI deployment: a powerful technology that could genuinely advance human creativity, packaged as a feature that mostly enables people to make worse versions of content that already exists. It's innovation in service of imitation, wrapped in a user interface that makes it feel like progress. The future is here, and it looks exactly like the present, but with bigger heads and more anime.