Microsoft has stated that it is not utilizing customer data from its Microsoft 365 apps for training its AI models. This clarification comes in response to reports circulating online in the last few weeks alleging that Microsoft required Word and Excel users to opt out of training the company’s AI systems.
The confusion stemmed from a privacy setting in Microsoft Office regarding “optional connected experiences” – a feature facilitating tasks like “search for online pictures” or “find information available online.” This setting is activated by default and does not explicitly mention AI training in the disclosure. Additionally, a Microsoft document from October 21st, 2024 listed several connected experiences in Office that “analyze your content” without specifically excluding AI training, adding to the confusion.
“In the M365 apps, we do not use customer data to train LLMs,” stated the Microsoft 365 X account to address the allegations. “This setting only enables features requiring internet access like co-authoring a document.” Microsoft’s communications head Frank Shaw also refuted the claims on Bluesky.
The incidents involving Adobe and Microsoft highlight a growing concern among individuals regarding their personal data being utilized by tech companies for AI training without explicit consent. This concern is understandable given that companies like Meta, X, and Google automatically opt their users into AI training and the extensive amount of online content being scraped for this purpose.
