Microsoft Clarifies AI Training: Your Office Docs Are Not Used

Microsoft Clarifies Data Usage in AI Model Training

Microsoft recently addressed confusion regarding the use of customer data from its Microsoft 365 applications for training AI models. Reports had been circulating that users of Word and Excel were required to opt out of having their data used for AI training, raising privacy concerns.

Understanding the Privacy Setting

The confusion primarily stemmed from a privacy setting in Microsoft Office known as "optional connected experiences." This feature is designed to enhance user experience by enabling functionalities like searching for online pictures or gathering information from the web. Notably, this toggle is activated by default, yet it does not make any mention of its implications concerning AI training.

Clarification from Microsoft

In response to the uproar, the Microsoft 365 X account stated, "In the M365 apps, we do not use customer data to train large language models (LLMs)." The account further clarified that the setting merely allows features that require internet access, such as co-authoring documents.

Insights from the Communications Department

Frank Shaw, Microsoft’s communications lead, took to Bluesky to debunk the rumors circulating online. This is not an isolated incident, as it echoes a similar situation faced by Adobe earlier this year.

Adobe’s Predicament

Adobe was also caught in the crossfire of privacy concerns when its user terms were misinterpreted, leading many to believe the company was using its users' work to train generative AI models. In response to this backlash, Adobe swiftly revised the wording in its terms of service to clarify the misunderstanding.

Growing Concerns About Data Privacy

The incidents involving Microsoft and Adobe signify a rising concern among users regarding the use of their personal data by tech giants for AI training. This apprehension is not unfounded, particularly considering companies like Meta, X, and Google that opt users into AI training by default. The extensive scraping of online content for such purposes only adds to the worries surrounding data privacy.

Conclusion

As users continue to express their concerns regarding privacy and data usage, it’s crucial for tech companies to maintain transparency and clear communication. Both Microsoft and Adobe have taken steps to clarify their data usage policies, but the conversation surrounding data privacy in AI remains a pertinent issue.