The summary feature of the chatbot relies on preprocessed movie data or existing captions and transcripts. Microsoft’s AI Copilot in the Edge website now has the capability to generate textual descriptions of movies. However, this feature is limited and works best with preprocessed videos or those with subtitles, according to Mikhail Parakhin, Microsoft’s CEO of advertising and web services.
Parakhin mentioned that the functionality of Edge Copilot depends on preprocessing the images or relying on existing subtitles. If a video lacks subtitles and has not been preprocessed, the feature will not work as intended. Essentially, Edge Copilot summarizes the text transcripts of videos rather than the videos themselves. This functionality extends to Microsoft 365’s Navigator, which can summarize Teams video meetings and calls for customer service representatives, provided that the audio is transcribed by Microsoft.
A recent demonstration by Pietro Schirano showcased Edge Copilot effectively summarizing a YouTube video about GTA VI. The user simply selected the “Generate movie description” option, and within moments, Copilot produced a detailed summary with highlights and timestamps.
While platforms like YouTube and Vimeo automatically generate transcripts and subtitles if users activate the feature, Parakhin indicated that Edge Copilot should work for most publicly available videos, such as those on YouTube.
Navigator exemplifies Microsoft’s ongoing competition in the AI landscape, particularly with Google. Recently, Google enhanced its Bard robot on YouTube to describe video content and extract specific information. Despite Microsoft’s advancements, Google’s Gemini update has faced challenges, with potential misrepresentations of the AI’s capabilities during demonstrations.
Parakhin has been transparent about Copilot’s development stages on social media, announcing updates like the integration of video information capabilities while on a flight.
