Preview users have noticed OneDrive’s AI-driven face recognition setting is opt-out, and can only be turned off ‘three times a year’

Is AI the future, or just an ecologically ruinous, dangerously huge financial bubble? That’s hard to say, but it’s certainly a part of Microsoft’s future; on top of pushing Copilot as a non-negotiable part of the Microsoft 365 suite and reportedly mandating its employees put AI to work, the tech giant is reportedly rolling out a new biometric collection setting that can only be turned off three times a year.

A Slashdot story yesterday relayed an editor’s experience when he noticed that, after uploading a photo from his phone’s local storage to Microsoft’s OneDrive cloud storage, the privacy and permissions page noted an opt-out “people section” feature wherein “OneDrive uses AI to recognize faces in your photos to help you find photos of friends and family.” When he went to turn it off, a disclaimer stated: “You can only turn off this setting three times a year.”

It doesn’t sound great, but users are speculating about what might motivate the restriction. Commenter on the Slashdot article AmiMoJo replied, “There might be a more benign reason for it. In GDPR countries, if you turn it off they will probably need to delete all the biometric data … if people toggle it too often, it’s going to consume a large amount of CPU time.” Regardless, the choice to make it opt-out rather than opt-in is sure to raise eyebrows among privacy advocates.

Speaking with Slashdot, Microsoft chose not to explain the reasoning behind the rule, but said “OneDrive inherits privacy features and settings from Microsoft 365 and SharePoint, where applicable.” The same article included a quote from a security activist from the Electronic Frontier Foundation, Thorin Klosowski, who said: “Any feature related to privacy really should be opt-in and companies should provide clear documentation so its users can understand the risks and benefits to make that choice for themselves.”

I can’t definitively say why the setting was rolled out this way, but it’s hard not to feel like this is yet another surreptitious means for AI to corrode privacy standards and encroach where it’s not always welcome. In a vacuum, it’s thorny—but paired with the nascent prominence of Copilot, LLM chatbots, Microsoft’s internal insistence on AI, and so on, it feels like another data point in a trend worthy of scrutiny.

About