On October 20, YouTube released its "Human Creativity and AI" report—the platform's first major study on why creators are using AI for video. The headline statistic: 72% of those surveyed have already used AI on their videos or images, and 61% want to use AI to insert themselves into the frame. YouTube's conclusion is surprising: creators are using AI not to save time, but to express themselves—a clear signal of which video formats are currently trending.
What the Numbers Say
The survey for YouTube was conducted by the agency NRG. A few important details to keep in mind: the respondents were all American, the sample size was small, and the study was commissioned by the platform itself, which is actively promoting its own AI tools.
| What | Data |
|---|---|
| Conducted by | NRG agency, commissioned by Google and YouTube |
| When | April 2026 |
| Where | USA |
| Who was surveyed | 694 people aged 14–44 who had uploaded a video online in the last 12 months |
| Used AI for videos or images | 72% |
| Want to use AI to insert themselves into videos | 61% |
| Of those who have posted AI videos, feature themselves | More than two-thirds |
The platform frames the report's main conclusion this way: generative video is driven by the "timeless human need for self-expression"—the same force that has always propelled online video. In other words, for creators, AI has become less a tool for speeding up editing and more a way to try on a new role.
The Most Popular AI Video Formats
Meeting Your Younger Self
The most prominent trend from the report: a creator takes a childhood photo, animates it with an AI tool, and "meets" their younger version—hugging them, talking to them, and showing them who they've become. YouTube explains the format's popularity by noting that the video gives viewers a ready-made template for complex emotions, and the comments section turns into a shared conversation, with people posting their own childhood photos in response.
From a practical standpoint, this is a game-changer for creators: a video like this is nearly impossible to shoot with a camera, yet it generates engagement, not just views.
Digital Costume Change
The second common scenario is placing yourself in a different profession, era, or scene: as a doctor, a pilot, a movie hero, or a game character. The report calls this a "digital costume change": the person remains themselves; only the world around them changes. This also includes filming "yourself" in cinematic scenes and creating entire fictional universes around a single character.
Hobbies Without Real-World Limits
The third group consists of those who don't run their channel for profit but simply pursue their passion. Fashion designers turn sketches into photorealistic runway shows, fans create cosplay for non-existent characters, and others visualize fanfiction. The report cites the example of a Minecraft creator named Kristufer, who films ASMR mukbangs in fictional, block-built restaurants.
Key Takeaways for Creators
- A human face is key. The report's stats are clear: of those who post AI videos, more than two-thirds feature themselves. A fully generated video without a human element is not what's working right now.
- AI is the prompt, not the plot. The most powerful formats from the report are built on personal stories: a childhood photo, your profession, your hobby. AI simply removes the "this is impossible to film" barrier.
- Optimize for comments, not just views. The "meeting your younger self" trend works because it prompts viewers to respond with their own photos. If your idea doesn't invite a viewer response, it needs more work.
- Check which tools are available to you. The report describes the behavior of an American audience with access to all of YouTube's AI features. Some services aren't directly available in Russia and Belarus, so it's best to build your own list of accessible tools.
YouTube's Rules: When to Label Your Video
One aspect the report doesn't cover is that you can face penalties for undisclosed AI use. YouTube's Help Center requires you to label content that is AI-generated or altered and looks realistic.
| What you did | Label required |
|---|---|
| A realistic scene of an event that never happened | Yes |
| A real person saying or doing something they didn't do | Yes |
| Altered footage of a real event | Yes |
| AI-generated music as the video's soundtrack | Yes |
| An obviously unreal world (e.g., a unicorn, a cartoon scene) | No |
| Beauty filters, color correction, or lighting adjustments | No |
| Automatic subtitles | No |
| Cloning your own voice | No |
The label is added during upload in YouTube Studio. In the video's details, find the question about altered content and select "Yes." YouTube states that for creators who repeatedly fail to disclose this, the platform may add the label itself and apply penalties, up to and including video removal and suspension from the YouTube Partner Program.
Another detail: according to Mediapost and Creator Handbook, in May 2026, YouTube made its facial recognition technology available to all eligible users over 18—it was previously only available to members of the Partner Program. It works similarly to Content ID: it finds videos where your likeness is used and allows you to submit a removal request. To enable it, you need to complete a one-time verification with an ID and a short video selfie.
What We Still Don't Know
- The sample is small. The 72% figure is based on just 694 people in the US and can't be extrapolated to countries like Russia, Belarus, and Kazakhstan, where available tools and user habits differ.
- There's a conflict of interest. The report was commissioned by YouTube, which sells its own AI features. Tubefilter points to this and notes the discontent among some creators regarding models being trained on their videos without consent and the flood of generic AI videos in search results.
- The report contains no financial data. It didn't measure how much money AI videos generate for creators or how they are monetized.
- The definition of "using AI" is unclear. Auto-subtitles and skin-smoothing filters are technically AI, and some respondents likely included them in their answers.
FAQ
Do I need to label a video if I used AI to insert myself into it?
If the scene looks realistic, yes, a label is required. If you're in an obviously fictional world or a cartoon style, no label is needed.
Can you monetize AI videos on YouTube?
Yes, there's no ban on using AI. However, YouTube requires videos to be original and transformative. Repetitive, mass-produced content from a single template is not eligible for monetization.
Will my channel be banned for AI videos?
Not for using AI itself. The risk comes from other issues: systematically failing to disclose AI use, using someone's likeness without consent, or posting mass-produced videos with no creative input.
What is the "meeting your younger self" trend?
A creator animates their childhood photo with AI and films a "meeting" with their younger self. YouTube cites this format as a prime example of how AI is being used for self-expression.
How many people actually use AI for video?
According to YouTube's survey, 72% of Americans aged 14-44 who have uploaded a video in the last year. This is not the percentage of all internet users and does not represent data for Russia.
Read More
- AI Search on YouTube: How 'Ask YouTube' Works and What It Means for Creators
- The YouTube Recommendation Algorithm: 83% of Creators Depend on It
- Does Instagram Reduce Reach for AI Videos? Mosseri's Answer and AI Labels
- How Many People Use Neural Networks in Russia: 55% and the Top 3 Services
If you're already creating short videos and want them to earn money, check out the "Content Rewards" section on Prime Oracles. Brands post briefs there, and creators can accept them with the "Accept Brief" button to get paid per view (CPV).
Source: YouTube Blog, October 20, 2026.
