Microsoft denies all allegations of taking user data from MS 365 Train AI Models applications

Recently, a number of Microsoft users have accused the company of using their data from Microsoft 365 applications, including Word and Excel, to train AI models. Users on social media have urged the tech giant to offer an option to opt out of the “connected experiences” feature. They claimed that this feature was used to train AI models. After this, Microsoft finally issued a statement to refute all these allegations.

A Microsoft spokesperson told Reuters: “These claims are untrue. Microsoft does not use customer data from Microsoft 365 business and consumer applications to train large underlying language models.” The spokesperson added that this feature enables features including co-authoring and cloud storage. The spokesperson also said it had nothing to do with how the company trains its large language models.

However, despite the statement, it seems that users are still not convinced whether their data is safe or not. Allegations and controversies continued to surface on social media platforms.

How did it all begin?

The issue arose from the “Connected Experiences” privacy setting in Microsoft Office, an optional feature that provides users with design suggestions, editing tips, data insights and other enhancements by analyzing Office content. While Microsoft detailed these features in a blog post, it didn’t explicitly confirm whether user data is used for additional purposes, such as AI model training. Notably, the feature is enabled by default, requiring users to manually disable it if they wish to opt out.

Concerns about data privacy in Microsoft Office were also highlighted in a post by nixCraft, a popular forum for Linux and Unix users on X. Microsoft issued a clarification from its official Microsoft 365 handle and wrote: “In M365 apps, we don’t use clients. data for LLM training. This setting only enables features that require Internet access, such as co-authoring a document.”