Microsoft Denies Using Customer Data to Train AI Models Amid Privacy Concerns

Microsoft Denies Using Customer Data to Train AI Models Amid Privacy Concerns

Microsoft declared that it does NOT use data collected from its Microsoft 365 applications such as Word and Excel to train AI, but explained their ‘connected experiences’ are different and users are in full control of their privacy settings.

Microsoft has strongly rejected recent claims that it trains its AI models with data from Microsoft 365 apps like Word and Excel. After posts on social media said that the company’s “connected experiences” feature, which is on by default, was being used to train AI, the company replied to these claims.

An official from Microsoft told Reuters that the claims were false and said, “These claims are false.” Microsoft does not train basic large language models with customer data from Microsoft 365 consumer and business apps.

The spokesperson made it clear that the “connected experiences” feature, which drives features like co-authoring and cloud storage, has nothing to do with how the company trains its AI.

Even though Microsoft gave an answer, worries about data privacy are still going around online. Some users are worried that their data could be used for AI training without their permission.

Microsoft also defended the idea saying that the “connected experiences” have been active by default since the time the feature was launched in April 2019. This function enhances the user experience due to features such as real-time grammar checks and web-based materials. The company stated that users would have 100 percent control over the privacy settings and the tool can be shut down at any time.

As the debate around Artificial Intelligence gathering people’s data and privacy continues, Microsoft aims to reassure its clients that it does not use their data to feed AI without their consent.

Leave a Reply

Your email address will not be published. Required fields are marked *

Microsoft AI Unveils LazyGraphRAG A Cost-Effective Solution for Scalable Graph-Enabled RAG Systems Previous post Microsoft AI Unveils LazyGraphRAG: A Cost-Effective Solution for Scalable Graph-Enabled RAG Systems