Microsoft Began Unauthorizedly Collecting Data From Word and Excel Documents for AI Training

The “Connected Experiences” feature in Microsoft Office, designed to analyze user-generated content, has reportedly transitioned to a new operational model. According to a post by social media user nixCraft on platform X, all user content is now transferred to an AI training dataset unless explicitly opted out. Microsoft has not yet commented on this claim.

By default, this setting allows Microsoft to use articles, artwork, and other documents opened in Office applications for AI training without requiring individual user consent. For those concerned about protecting their intellectual property or confidential information, it’s advisable to take action. Users can opt-out by adjusting the settings, though on Windows devices, this option is buried several clicks deep within the File menu. The process involves unchecking the “Enable optional connected experiences” box, which is enabled by default.

On a Windows PC, the steps are as follows:

  1. Go to File > Options.
  2. Select Trust Center > Trust Center Settings.
  3. Navigate to Privacy Options > Privacy Settings.
  4. Find the Optional Connected Experiences option and uncheck the corresponding box.

Microsoft’s approach reflects a broader trend in the tech industry, where AI developers actively seek materials to train their models. While all AI systems rely on human-created content for learning, doing so without explicit user consent raises ethical concerns.

The company has yet to confirm or deny whether data from user-created Excel and Word documents are being used to train its AI models. However, Microsoft’s website includes a document titled Microsoft Services Agreement, which outlines user permissions.

One clause states:

“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 Microsoft a worldwide, royalty-free license to use intellectual property in your content. This includes copying, storing, transmitting, reformatting, displaying, and distributing your content in the services through communications.”

This clause indicates that user content may be utilized for various purposes, further fueling concerns about data privacy and usage transparency.

Scroll to Top