Is Microsoft Using Your Word Documents to Train AI?

People are increasingly worried about how AI affects them. Will our jobs become obsolete? Are we constantly being tracked online and in the real world? Most recently, people have raised concerns about whether our creative and professional work is being used to train large language models.

Over the past couple of weeks, rumors have surfaced that Microsoft is scraping customers’ Word documents and Excel spreadsheets to train its AI systems. These claims arose after Microsoft introduced a new feature in October called Connected Experiences, a tool buried within its programs’ Options menus that is activated by default unless you opt out. It contains three checkboxes—”Turn On Experiences That Analyze Your Content,” “Turn On Experiences That Download Online Content,” and “Turn On All Connected Experiences.”

One could be forgiven for being suspicious. After all, wording like “turn on experiences that analyze your content” is enough to set off alarm bells in the minds of anyone who is justifiably vigilant about how AI behaves. The vague wording on Microsoft’s Connected Experiences explainer page also doesn’t put these concerns to bed, nor does the company’s Privacy Statement, which talks about building, training, and improving “the accuracy of our automated methods.” Moreover, on its Copilot page, Microsoft says it “uses data from Bing, MSN, Copilot, and interactions with ads on Microsoft for AI training.” But does it use your creative or professional work in the Microsoft 365 apps to do the same?

We contacted Microsoft to find out. In response, a Microsoft spokesperson said, “Microsoft does not use customer data from Microsoft 365 consumer and commercial applications to train large language models. Additionally, the Connected Services setting has no connection to how Microsoft trains large language models.”

Adding more context, the spokesperson continued, “The Connected Services setting is an industry standard setting that enables features that require an internet connection. Connected Experiences play a significant role in enhancing productivity by integrating your content with resources available on the web.” They also provided specific examples of how Connected Experiences can help you with your work:

• “Providing relevant suggestions: When you’re writing a document in Word, it can suggest relevant information or images from the web.
• Enhancing collaboration: Features like real-time co-authoring and cloud storage integration make it easier to work with others.
• Improving productivity: Tools like Editor in Word can offer grammar and style suggestions by analyzing a vast amount of data.”

The company also posted on X to confirm that it does “not use customer data to train LLMs.”

Either way, if you still feel uneasy about the Connected Services feature, you can opt out by opening the File menu in Word or Excel, selecting “Options,” navigating to “Trust Center,” and selecting “Trust Center Settings.” From there, open “Privacy Options,” navigate to “Privacy Settings,” and uncheck the three options listed in the Connected Experiences section.

Source: Microsoft