Microsoft began to unauthorizedly collect data from Word and Excel documents for AI training – it’s not easy to refuse

The Connected Experiences set of functions present in the Microsoft Office suite, designed for analyzing materials created by users, has been transferred to a new operating mechanism – all content is transferred to an array for training artificial intelligence, unless explicitly stated otherwise, noted a user of the social network X under the nickname nixCraft. Microsoft has not yet commented on this information.

Image source: BoliviaInteligente / unsplash.com

The default setting gives Microsoft the right to use articles, artwork, and other documents opened in Office applications to train AI without asking for user consent in each case. Therefore, anyone concerned about protecting their intellectual property or confidential information is advised to take action. The user can refuse this behavior of office programs – to do this, you need to find the corresponding option in the settings, but in the case of a PC running Windows, it is seven clicks deep in the “File” menu.

Microsoft’s approach reflects a general trend in the tech industry: AI developers are actively looking for materials that can be used to train models – all of which are trained on human-created content, but doing so without the explicit consent of the consumer may not be entirely ethical. The company has not yet officially confirmed or denied that it uses data from Excel and Word documents created by Office users to train the AI.

At the same time, a document called “Microsoft Services Agreement” is posted on the company’s website. “To the extent necessary to provide the services to you and others, protect you and the services, and improve Microsoft products and services, you grant Microsoft a worldwide, royalty-free license to use the intellectual property associated with your content, such as copying, storing, transmitting, reformatting, displaying and distributing your content across services using communications media,” reads one of the paragraphs of the document.