Is Microsoft Using Users Word Documents to Train AI?

People are increasingly concerned about how AI will affect their lives. Will our jobs become obsolete? Will we be constantly tracked online and in the real world? Similarly, many tech companies, large and small, are struggling to find quality data to train their AI models, and this is where the controversy comes in.

Over the past few weeks, rumors have been circulating that Microsoft is collecting customers’ Word documents and Excel spreadsheets to train its own AI systems. These claims come after Microsoft introduced a new feature in October called Connected Experiences, a tool hidden in the Options menu of Office apps that is enabled by default unless users opt out. The feature consists of three component options:

• Turn On Experiences That Analyze Your Content
• Turn On Experiences That Download Online Content
• Turn On All Connected Experiences

Translation:

• Turn on Your Content Analytics Experience
• Enable Online Content Download Experience
• Turn on all Connected Experiences".

Without going into too much detail, the names of these options suggest some potential “risk” to personal data security and privacy. The vague wording on Microsoft’s Connected Experiences explainer page does little to address these concerns, nor does the company’s Privacy Statement, which talks about building, training, and improving “the accuracy of our automated methods.”

Furthermore, on the Copilot site, Microsoft says they “use data from Bing, MSN, Copilot, and Microsoft ad interactions to train the AI.” But will they use your creative data in Microsoft 365 apps to do the same?

In a press release, a Microsoft spokesperson said, " Microsoft does not use customer data from Microsoft 365 consumer and commercial apps to train large language models. Additionally, Connected Services settings have no bearing on how Microsoft trains large language models. "

Microsoft also emphasized that "Connected Services is a standard setting that enables features that require an internet connection. Connected experiences play a key role in enhancing productivity by integrating user content with resources available on the web." The company also provided specific examples of how Connected Services can support users in their work:

• Provide relevant suggestions: When you're drafting a document in Word, AI can suggest relevant information or images from the web.
• Improved collaboration: Features like real-time collaboration and cloud storage integration make it easier to work with others.
• Improved productivity: Tools like Editor in Word can make grammar and style suggestions by analyzing large amounts of data.

Microsoft also posted an official post on social network X confirming that they "do not use customer data to train LLMs".

That's a promise from Microsoft, if you're still uncomfortable with 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 ". Here, open " Privacy Options ", navigate to " Privacy Settings " and uncheck the three options listed under Connected Services .