Microsoft’s AI Copilot within the Edge browser has rolled out an innovative feature: the ability to craft text summaries for videos. However, this function comes with a catch—it solely operates on pre-processed videos or those already equipped with subtitles. As clarified by Mikhail Parakhin, Microsoft’s CEO of advertising and web services, the AI relies on pre-processing or existing subtitles to generate these video summaries.
Effectively, Edge Copilot doesn’t directly summarize videos; instead, it works on text transcripts extracted from the videos. The AI extends this functionality across Microsoft 365, summarizing Teams video meetings and calls but requires prior transcription by Microsoft. Similarly, Copilot on Microsoft Stream offers summarization but necessitates a written transcript.
The buzz ignited when designer Pietro Schirano shared a recording of Edge Copilot summarizing a YouTube video on the GTA VI trailer. While seemingly seamless in this instance, Copilot's reliance on pre-processed content was highlighted. Parakhin responded that this feature should work for most publicly available videos, although pre-processing remains a prerequisite.
Despite Copilot's stride in text summarization, it faces stiff competition in the generative AI arena, notably from Google. Last month, Google upgraded its Bard chatbot’s YouTube extension, enabling video content summarization. However, this week's Gemini update by Google faced scrutiny for potential misrepresentation of AI capabilities and inconsistencies in content accuracy.
Parakhin’s transparency about Copilot’s evolution echoes in his social media updates, revealing ongoing efforts to enable Copilot to harness video information even while on a flight. Microsoft's foray into generative AI through Edge Copilot marks its quest to innovate amid evolving competition.