YouTube: 72% of uploaders use AI on videos or images

A YouTube-commissioned survey of nearly 700 video posters aged 14-44 found 72% use AI on videos or images, with many inserting themselves into content. The trend signals both creative opportunity and misinformation risk.

What happened

YouTube published a new report in its Culture & Trends series looking at how creators are integrating artificial intelligence creation tools into their workflows. The platform commissioned NRG to survey nearly 700 people aged 14 to 44, who have posted a video online in the last 12 months, regarding their use of AI tools.

The data found that despite various concerns around the use of AI tools, more creators are integrating AI into their workflow in order to create new types of content. YouTube said: "A surprising 72% of uploaders (surveyed video posters aged 14–44) say they use AI on videos or images." In a U.S. study of people aged 14–44, 61% of people who have posted videos in the past 12 months said they were interested in using AI to insert themselves into video content.

According to YouTube, this self-insertion was the most popular usage of its AI tools among the survey group, with more than two-thirds of people who’ve posted AI videos indicating that they featured themselves in their content. YouTube described the appeal: "Adding yourself to AI content lets you play at being the director of your own life — making yourself the center of a comedy skit, or turning your family history into a period piece."

Key facts

• YouTube commissioned NRG to survey nearly 700 people aged 14 to 44 who have posted a video online in the last 12 months.
• 72% of uploaders surveyed video posters aged 14–44 say they use AI on videos or images.
• 61% of people who have posted videos in the past 12 months said they were interested in using AI to insert themselves into video content.
• More than two-thirds of people who’ve posted AI videos indicated that they featured themselves in their content.
• Reported examples include digital costume changes, inserting themselves into video game scenes, fashion designers bringing concepts to life, and creators imagining entire cinematic worlds.

Our analysis

The 72% usage figure suggests that generative AI has already moved from an experimental novelty to a common step in video creation for this group. The emphasis on inserting creators into scenes—more than two-thirds of AI video posters featured themselves—points to AI being used less as a faceless production tool and more as a form of identity play. For social teams, that likely means audience expectations are shifting: viewers may increasingly expect creators to appear inside imagined worlds, historical settings or game scenes rather than only in traditional filmed formats.

However, the same personalization creates obvious risks. YouTube’s report acknowledges that adding a generative version of another person into a clip could lead to misinformation or misleading depictions. A brand or community operator running creator campaigns will likely need to distinguish between consented self-insertion and unauthorized likeness use, especially as this type of output grows. The report’s reference to balancing the influx of AI slop with a much smaller percentage of valuable storytelling expansions suggests that discovery algorithms and community guidelines may increasingly reward polished, intentional use over mass-generated filler.

What is different from earlier AI adoption is that the value is shifting toward creator identity and world-building rather than only editing efficiency. YouTube’s language about “AI auteurs” and cinematic universes indicates the platform sees a new class of creator emerging, which may affect partnerships, content formats and moderation priorities.

What it means for operators

• Treat AI self-insertion as a creative format: plan templates or prompts that let creators place themselves in on-brand scenes, rather than only requesting standard UGC videos.
• Set likeness and disclosure rules before commissioning AI content: require consent for any real person inserted into a clip and document which parts are synthetic, because misleading depictions are called out as a risk in the report.
• Use the 72% usage finding to pitch AI-assisted video content to stakeholders or clients as a mainstream creator workflow, not a fringe trend.
• Separate high-effort AI storytelling from low-effort slop in campaign briefs and creator selection; the report frames the challenge as balancing the influx of AI slop with a much smaller percentage of valuable storytelling expansions, so quality signals may become a competitive advantage.