Microsoft’s OneDrive Introduces AI-Powered Photo Grouping Feature
In a move that is sure to stir conversations around privacy and user control, Microsoft is rolling out an AI-driven feature in OneDrive designed to enhance photo organization by recognizing faces. This innovative capability allows users to group images based on identified individuals, although it comes with certain limitations that may raise eyebrows among privacy advocates.
Currently, a select group of users, including this writer, have been enrolled in a mobile preview of the feature, which has already begun appearing on various devices. Microsoft has indicated that the full rollout is forthcoming, but the implications of such technology are already being scrutinized.
The functionality requires users to manually identify faces in their photos, a process that, while not groundbreaking, sets the stage for OneDrive to curate collections based on these identifications. However, the feature’s design raises questions about user autonomy. According to Microsoft’s documentation, users can only toggle the People section on or off three times within a year, a stipulation that has yet to be clarified by the company.
In response to inquiries about this limitation, Microsoft has remained tight-lipped, leaving many to wonder why such a seemingly arbitrary cap exists. The question arises: why not allow users the flexibility to adjust their settings as they see fit? This ambiguity may contribute to unease among those who prioritize their privacy.
Moreover, Microsoft has acknowledged that certain regions mandate user consent before processing photos, hinting at potential regulatory scrutiny. For instance, it took until July 2025 for Microsoft to launch its Recall feature in the European Economic Area (EEA), largely due to data processing concerns. This historical context adds another layer of complexity to the current rollout.
Should users choose to disable the facial grouping feature, all associated data will be permanently erased within 30 days. However, the lack of clarity surrounding the definition of a “year” in this context—whether it begins with the initial setting change or the start of face identification—further complicates the user experience.
While Microsoft assures users that it does not utilize facial scans or biometric data for its AI training and that grouping data remains private, the concept of using images in this manner may still evoke discomfort among some customers. As the feature remains in its preview stage, the dialogue surrounding its implementation and user implications is likely to continue evolving.
