Google integrates advanced AI video analysis into Ask YouTube

Photograph: Nwz/Shutterstock. YouTube logo: Source: YouTube. Exposure/shadow correction, crop, and separate official-logo band. — Image credit: Search Engine Journal

Google announced Tuesday it is integrating new ‘agentic video understanding’ technology into Ask YouTube, allowing its Gemini AI model to analyze specific video segments more efficiently and accurately for user queries.

This enhancement will power Ask YouTube directly on the video watch page in the coming months, promising significant improvements in how the platform processes and responds to questions about video content.

The new system enables Gemini to selectively load specific parts of a video, dynamically adjust frame rates, and determine whether to include frames, audio, or transcripts for particular segments, according to Google.

Google stated the technology reduces token usage by up to 88 percent and lowers analysis costs by as much as 66 percent compared to static video processing.

Accuracy for video understanding also improves by up to 7 percent with the new agentic approach, the company reported.

Brooke Osmundson, a Google spokesperson, said the technology marks a substantial step forward in making video content more accessible and searchable through AI.

Alphabet CEO Sundar Pichai has previously highlighted the importance of advanced AI in enhancing user experience across Google’s diverse product portfolio.

Developers can already access this feature through the Gemini API in Google AI Studio and the Gemini Enterprise Agent Platform.

The watch-page version of Ask YouTube operates independently from the search bar version, which currently provides video summaries with cited sources.

Source: Search Engine Journal