AI search is not just changing where people look for answers. It is changing whose answers get treated as useful.
A new Jellyfish analysis shared with ADWEEK found that YouTube creator content appears in more than one in four answers generated by AI assistants. That already makes YouTube a serious source layer for AI discovery. But the number gets more interesting in categories where users are closer to making a decision.
In consumer electronics and financial services, Jellyfish found that YouTube creator content shows up in nearly one in two AI assistant responses. In CPG, including household cleaning product prompts, the agency found that more than one million unique YouTube videos are cited daily by AI assistants in the U.S. alone.
The research is not small. Jellyfish, which is part of The Brandtech Group, analyzed 27 million responses across seven AI assistants, including Claude, ChatGPT, Gemini, DeepSeek, Meta AI, and Perplexity. Its finding is blunt: independent, niche creators are beating brand-owned content and celebrity influencers inside AI-generated answers.
Creators are becoming the sources AI trusts
For years, YouTube has been the place people went when they wanted to see something explained, tested, reviewed, unboxed, compared, repaired, or demonstrated. The platform’s creator model was built around that behavior long before AI search became a marketing concern. A 12-minute product review, a niche comparison video, or a practical how-to may not look like a polished brand asset. But it is exactly the kind of content that can contain the context an AI assistant needs to answer a question.
That helps explain why the Jellyfish data points toward creators rather than celebrities. Celebrity influence is useful when attention is the goal. AI search is different. It rewards answerability. A creator with a small but focused audience may produce more detailed, category-specific content than a global ambassador with a glossy campaign post.
The reported preference for videos longer than 10 minutes also fits the behavior. Longer YouTube videos often include setup, comparison, caveats, product use, personal experience, and follow-up comments from the creator. That is not always elegant. But it gives machines more substance to parse, summarize, and cite.
This is where the shift gets uncomfortable for brands. Many brands have invested heavily in owned content, SEO pages, influencer campaigns, and paid social assets. But if AI assistants are pulling from the content that best answers user prompts, then controlled brand messaging may be less competitive than credible creator explanation.
That does not mean brands should abandon owned content. It means owned content has to work harder. A product page that repeats claims may be less useful to an assistant than a creator video that demonstrates the product against three alternatives. A campaign film may build fame, but it may not answer the question someone asks ChatGPT before buying.
For creators, this adds a new layer to YouTube’s value. The audience is no longer only the people who click play on YouTube. It may also be the AI systems that use creator videos to shape answers elsewhere. Visibility can now happen without the viewer ever landing on the original video first.
The search playbook is getting rewritten
The practical implication is simple: discoverability is moving beyond the search results page. Brands used to optimize for Google rankings, retail search, social feeds, and platform algorithms. Now they also have to think about whether their category knowledge is showing up inside AI-generated answers.
That is not just a technical SEO problem. It is a content problem. The Jellyfish findings suggest that AI assistants may favor content that behaves like evidence: specific, explanatory, comparative, and useful in moments of intent. In other words, the kind of content many niche YouTubers have been making for years.
This creates a new kind of creator advantage. The most valuable creator may not be the one with the biggest reach. It may be the one whose content becomes part of the answer set when a consumer asks, “Which laptop should I buy?”, “What is the safest credit card for travel?”, or “Which cleaning product actually works on this surface?”