YouTube Turns Search Into a Chatbot — and Publishers Should Worry

YouTube will let viewers build custom recommendation feeds by typing plain-language prompts, the company announced Wednesday at its Made on YouTube event. The feature, rolling out across desktop, TV and mobile in the US next month, creates dedicated homepage tabs for topics a user asks for — photography, college football, or anything else.

The move pushes the platform's machine learning, which has shaped feeds for years, into generative territory: viewers will effectively prompt YouTube the way they prompt a chatbot, and the platform will assemble a personalized feed in response.

YouTube is also expanding Ask YouTube, its experimental AI search tool. Ask a question about a product — the announcement used the hypothetical of a new iPhone Duo — and the tool compiles information from videos and outside sources to build comparison tables and other structured answers. Users can click a surfaced video and keep asking questions on the video's watch page.

The mechanics mirror how chatbots search the web, synthesize and cite. That parallel carries a business risk that YouTube did not fully dismiss: just as generative AI has drained traffic from digital publishers, AI-enhanced video search could let viewers get answers without clicking through to original content.

YouTube's response to that concern is quantitative. "Our goal is to connect viewers with the right video faster. With Ask YouTube, we expect watch time to remain steady or improve," a YouTube spokesperson told CNET. The spokesperson added that earlier AI features, chapters and AI-generated summaries, "showed no negative impact on watch time; instead, they helped retain viewers." Videos cited in Ask YouTube's answers will "prominently display the video title and channel name."

For Google, which owns YouTube, this is the continuation of a multi-year project to embed AI across the platform. The company has integrated tools powered by its Gemini models, and at Made on YouTube it said it is adding Gemini as a content creation assistant in Shorts and the YouTube Create app. Creators have already used AI to edit and produce videos at scale — a flood of material critics call AI slop — while the platform's AI effort now centers on personalization for both viewers and creators, aimed at keeping users watching and monetizing content.

The child safety context looms over the personalization push. A landmark court decision this year found that Meta and Google designed their sites to be intentionally addictive to children and teens, harming their mental health. YouTube disputes the ruling's application to itself, arguing it is a streaming platform, not a social media platform. Child safety researchers and advocates have warned about platforms adding more AI features; Reuters' 2025 investigation found Meta's AI chatbots on Instagram engaged in sexual and romantic roleplay with children.

Creator tools

YouTube announced a dedicated format for episodic content called Shorts series. Creators can designate videos as part of a series with season and episode numbers, and the series will live on a creator's page as a playlist. The tool targets the growing investment in serialized content, such as Sydney Jo's The Group Chat series and microdramas. At launch, access is limited to creators in YouTube's partner program, on web, mobile and TV.

The platform is also giving creators two new personalization levers. Creators will be able to test multiple thumbnails and preview clips per video; YouTube will serve the variants to real viewers and match each viewer with the version the algorithm predicts they are most likely to click. That kind of individual-level content matching is a core promise of AI for digital marketers.

Auto dubbing — real-time audio translation — is coming to livestreams, starting with English and Spanish, with more languages planned. And YouTube is adding a new affiliate partnership with Amazon, expanding how creators can monetize.

The throughline of the event is clear: YouTube is applying generative AI to both ends of its marketplace, making discovery conversational for viewers and targeting automatic for creators. Whether watch time holds, as the company predicts, will determine if the publisher-style traffic collapse that hit text media arrives on video platforms too.