YouTube has started rolling out a new “Ask” button powered by Gemini AI. The feature allows viewers to generate instant summaries, pull out key points, and ask specific questions about the video they’re watching, all without leaving the app.
Powered by Google’s multimodal Gemini AI, the new feature goes beyond simple transcript analysis. It understands what’s being said in the video, what appears on screen, and even the context of visuals or captions to deliver accurate and meaningful answers. When a viewer types a question, Gemini analyzes the video and responds with context-rich information, often including direct timestamps that take users straight to the part of the video that answers their query.
For example, someone watching a tech review can ask for the “main takeaways” or “price details,” and Gemini will give a short summary or point them to the exact segment where those details appear. It can also clarify visual data, like a chart or infographic, and even suggest follow-up questions to help viewers dig deeper into the topic.
The Ask button currently appears next to eligible videos for signed-in users, primarily on the YouTube mobile app. However, the company says it will expand to desktop users over time. For now, Gemini only draws information from YouTube’s ecosystem, and it doesn’t fetch data from the wider web, ensuring that responses remain tied to the actual video content.
The rollout began in October 2025 and is being expanded gradually across regions and devices. YouTube is refining the experience based on user feedback before making it widely available.
