Microsoft accused of using online Word and Excel content for AI training, unless you opt-out

Microsoft Office, which includes popular productivity apps like Word and Excel, has been accused of scraping user-created content to train its internal AI systems. The feature, called 'Microsoft Connected Experiences,' is enabled by default, and to opt-out, you need to navigate various settings and options.

3

Linux expert @nixcraft on X called 'Connected Experiences' "unethical" because the feature's name doesn't convey that it's all about AI training. Microsoft's knowledge base states that connected experiences are "designed to enable you to create, communicate, and collaborate more effectively" by analyzing your content. There's no mention of AI or data being used for Copilot features or to train Microsoft's AI models.

Various media publications have picked up on these allegations, and Microsoft has responded to them using the official @Microsoft365 account on social media, denying the claims.

"We do not use customer data to train LLMs," Microsoft said. According to the company, 'Microsoft Connected Experiences' isn't about generative AI tools or training. "This setting only enables features requiring internet access like co-authoring a document."

However, as Tom's Hardware points out, Microsoft 365 and Microsoft Connected Experiences are covered by Microsoft's Services Agreement, which gives the company full intellectual property rights to all of your content.

"To the extent necessary to provide the Services to you and others, to protect you and the Services, and to improve Microsoft products and services, you grant to Microsoft a worldwide and royalty-free intellectual property license to use Your Content, for example, to make copies of, retain, transmit, reformat, display, and distribute via communication tools Your Content on the Services," the agreement reads.

With web-based apps and services, social media, search engines, and more, several major players in the tech and AI space have been accurately called out for scraping user data to train AI. The CEO of Microsoft AI, Mustafa Suleyman, recently said that any content on the web "is fair use" for AI training, reproduction, and more. Everything else, which includes data that people actively opt out of sharing or checking a 'do not scrape' box, is a 'gray area.' Not only that, but it's up to the courts to decide what data is off-limits.

Back to Microsoft Connected Experiences, the company has stated that it is not using your Word documents to train AI models. Still, if you want to opt-out, you need to go to File > Options > Trust Center > Trust Center Settings > Privacy Options > Privacy Settings > Optional Connected Experiences > Uncheck the box: "Turn on optional connected experiences."