Google expands AI integration into YouTube with conversational search and Gemini Omni

Google has introduced 'Ask YouTube', an artificial intelligence-driven conversational search feature that allows users to pose complex queries and follow-up questions across both Shorts and long-form videos. The tool is currently available to Premium subscribers in the United States on desktop devices as part of a testing phase for new capabilities. This launch represents a significant expansion of Google’s search experience, moving AI integration beyond standard text-based queries into the video ecosystem.

The new feature enables users to refine their searches through iterative questioning, with examples including requests for tips on teaching a child to ride a bike or finding creator reviews of cozy games. YouTube compiles results from both short and long-form content to generate responses. Availability is currently restricted to US-based Premium subscribers on desktop, indicating a controlled rollout rather than a broad public release.

In parallel with the search update, Google is integrating its new Gemini Omni video model into YouTube Shorts Remix and the YouTube Create app. According to a press release from the company, the model is designed to better understand user intent to create more consistent storytelling while handling complex video and audio adjustments behind the scenes. YouTube described this integration as a way for users to create and build on each other’s imagination through remixing.

The company is also expanding its likeness-detection tool to creators aged 18 and older to mitigate the risk of deepfakes. The tool functions as a removal mechanism, allowing creators to request the deletion of videos where they are misrepresented in AI-generated content. This expansion comes as the platform seeks to address growing concerns around synthetic media, although the effectiveness of the measure remains to be seen given its recent broadening.

This move places YouTube in competition with other technology firms such as Meta and OpenAI, which have faced mixed reception regarding AI integration in short-form video. OpenAI previously sunsetted its social app Sora, while Meta’s efforts have drawn varied responses. YouTube’s approach appears characterised as less front-and-centre compared to these competitors, embedding AI tools more subtly into existing workflows rather than promoting them as standalone products.

The long-term viability and user adoption of 'Ask YouTube' and the Gemini Omni integration are not yet established, given the current limited availability. Furthermore, claims regarding Gemini Omni’s ability to deliver meaningful storytelling are based on promotional assertions from YouTube and should be treated as such rather than independently verified performance metrics. The likeness-detection tool is a reactive measure for misrepresentation after it occurs, rather than a preventative filter against deepfakes.