Windows previews ‘Intelligent Media Search’ that transcribes videos and audio

Microsoft is trialing a new Windows feature called “Intelligent Media Search” that transcribes media files into text that can be processed by AI. 

The feature would turn any spoken words in audio or video files on a Windows computer into text. This text output would be saved to a data file which could then be accessed by an AI model. 

Users who allow the feature would be able to conduct text searches for any words appearing in their videos. A content creator, for example, with an expansive library of videos stored on their device could use the “Intelligent Media Search” feature to find a video featuring a specific line or to organize all files featuring a specific set of words or phrases. 

According to a report from Tom’s Hardware, the feature is merely being trialed for select members of the Windows Insiders group — users who’ve opted in to features which may not yet be ready for public use. This means there’s no guarantee that current or future versions of Windows will actually receive the new feature.

Privacy concerns

The feature comes as Microsoft remains under fire from privacy advocates worried about its upcoming “Recall” feature that, essentially, allows Microsoft to track everything you do on your device.  

Early reports indicated that Recall would be a mandatory feature for Windows users going forward but the company has subsequently stated that it would be optional. 

The Intelligent Media Search feature would also be optional for Windows users. Per Tom’s Hardware, users who’ve tried it so far were required to consent to both the AI model’s download and its access to their files. 

Related: Area 51 for AI? US gov secures early access to cutting edge models