Microsoft’s AI Copilot, a feature in the Edge browser, has recently added a new capability of generating text summaries of videos. This feature has the potential to save users time by providing condensed versions of videos. However, the ability of Edge Copilot to summarize videos is still quite limited and comes with certain prerequisites.
According to Mikhail Parakhin, Microsoft’s CEO of advertising and web services, in order for Edge Copilot to work, the video needs to be pre-processed. If the video contains subtitles, Copilot can fallback on them to generate the summary. However, if the video lacks subtitles and hasn’t been pre-processed yet, the feature won’t work. This means that Edge Copilot does not actually summarize the videos themselves, but rather the text transcripts of the videos.
The capabilities of Copilot extend beyond the Edge browser. It can also summarize Teams video meetings and calls for customer service agents within Microsoft 365. However, similar to the Edge browser, the audio needs to be transcribed first by Microsoft in order for Copilot to perform its summarization function effectively. Additionally, Copilot on Microsoft Stream can summarize any video, but users are required to generate a written transcript for it to work.
A video posted by designer Pietro Schirano showcased Copilot successfully summarizing a YouTube video about the GTA VI trailer. By pressing the “Generate video summary” button in the Copilot sidebar, the user in the recording received a summary within seconds, complete with highlights and timestamps. However, it’s important to note that many platforms, including YouTube and Vimeo, already have the capability to automatically generate transcripts and subtitles if users enable the feature.
When asked if Edge Copilot would work for most publicly available videos, Mikhail Parakhin responded, “Should work for most videos.” This indicates that Microsoft believes the majority of videos, such as those on YouTube, do not require pre-processing. However, it remains to be seen how well Copilot can handle a wide range of videos without subtitles or pre-processing.
Competition with Google
Microsoft is in a race with Google and other companies in the generative AI space. Last month, Google upgraded its YouTube extension for the Bard chatbot, enabling it to summarize video content and extract specific information from it. However, Google’s recent Gemini update has faced issues, with potential misrepresentation of the AI’s capabilities in a demo and occasional inaccuracies.
Mikhail Parakhin has been transparent about the various stages of Copilot’s development on social media. This openness reflects Microsoft’s commitment to refining and improving the feature over time.
While Microsoft’s AI Copilot shows promise in generating text summaries of videos, its capabilities are currently limited to pre-processed videos or those with subtitles. The integration with Microsoft 365 expands its functionalities but relies on accurate transcription of audio. As the generative AI race between Microsoft and competitors like Google continues, it will be interesting to see how Copilot evolves and addresses the limitations identified.
Leave a Reply