YouTube videos can’t be summarized by Microsoft’s Edge Copilot AI
3 mins read

YouTube videos can’t be summarized by Microsoft’s Edge Copilot AI

Summarizing video data or existing transcripts and subtitles is done by the chatbot.

Microsoft’s AI Copilot in the Edge browser now has a new feature that can create written summaries of videos. This feature has some limitations—it works best with videos that have subtitles or have been pre-processed in a certain way. According to Mikhail Parakhin, who is the CEO of advertising and web services at Microsoft, the time-saving feature is still somewhat restricted in its capabilities.

MSPowerUser noticed that Mikhail Parakhin, the CEO of advertising and web services at Microsoft, mentioned that for the new feature to function, the video needs to undergo a pre-processing step. If the video already has subtitles, the feature can still generate a summary using them. However, if the video lacks subtitles and hasn’t been pre-processed, then the feature won’t work as intended.

Simply put, Edge Copilot doesn’t directly summarize videos; instead, it summarizes the text transcripts associated with the videos. This functionality extends beyond Edge, covering Microsoft 365 features like summarizing Teams video meetings and customer service calls. However, for these summaries to be generated, the audio content must first be transcribed by Microsoft. The same applies to Copilot on Microsoft Stream, where it can summarize any video, but users must first provide a written transcript.

The discussion began when designer Pietro Schirano shared a screen recording of Edge Copilot summarizing a YouTube video about the GTA VI trailer. The video showed Copilot performing well, with the user pressing the “Generate video summary” button in the Copilot sidebar. Within a few seconds, Copilot produced a summary of the video, including highlights and timestamps.

Many platforms such as YouTube and Vimeo offer automatic transcript and subtitle generation if users enable this feature. When TheOrcTech asked Parakhin on X whether we could assume most publicly available videos (like those on YouTube) weren’t pre-processed, he replied, “Should work for most videos.”

YouTube videos can't be summarized by Microsoft's Edge Copilot AI

Copilot is part of the ongoing competition in generative AI between Microsoft and Google, among others. Google recently upgraded the YouTube extension for its Bard chatbot, allowing it to summarize video content and extract specific information. Google also announced a significant update to Gemini this week, though it has faced some challenges, including potential misrepresentation of AI capabilities in a demo and inaccuracies in its information.

Parakhin has openly shared various stages of Copilot’s development on social media. While on a plane on Tuesday morning, he posted on X about adding the ability for Edge Copilot to use information in videos during the flight.

Leave a Reply

Your email address will not be published. Required fields are marked *