No, Microsoft isn’t using your Office docs to train its AI

Microsoft says it isn’t utilizing customer information from its Microsoft 365 apps to train its AI models. The explanation addresses reports circulating online successful the past fewer weeks claiming Microsoft required Word and Excel users to opt retired of training the company’s AI systems.

The disorder arose from a privateness mounting successful Microsoft Office that toggles “optional connected experiences” — a characteristic that helps users “search for online pictures” aliases “find accusation disposable online,” according to Microsoft. This toggle is switched connected by default, and fails to mention AI training successful the disclosure. Similarly, a Microsoft learning archive posted connected October 21st, 2024 seems to person contributed to the disorder by describing a agelong database of connected experiences successful Office that “analyze your content” without explicitly excluding AI training.

“In the M365 apps, we do not usage customer information to train LLMs,” the Microsoft 365 X relationship said, responding to claims. “This mounting only enables features requiring net entree for illustration co-authoring a document.” Microsoft’s communications caput Frank Shaw also chimed successful connected Bluesky to debunk the claims. 

The Adobe and Microsoft incidents propose that group are progressively concerned pinch their individual information being utilized by tech companies to train their AI models without definitive permission. It’s an understandable interest given companies for illustration Meta, X, and Google opt their users into AI training by default, and the immense quantities of online contented being scraped for that purpose.