Microsoft Research Focused on Demographic Data, Not PII

Erik Bratt, Sr. Communications Manager, here ...

At Microsoft, we’re fortunate to have some of the world’s most renowned technology scientists working on innovative research projects based on an open academic model. Microsoft Research is dedicated to conducting both basic and applied research in computer science and software engineering.

A recent research paper published by Microsoft Research Asia focused on analyzing the ability to predict users' gender and age, referred to as demographic information, from a list of Web sites they had visited. The project is based on a well-known machine-learning algorithm that has been in general use for many years. Click here to view the actual research.

There has been some confusion lately about what this research actually involves and its implications for privacy, and we want to provide some clarity.

First, we want to be clear that no personally identifiable information was used in this research. The researchers’ analysis was based on anonymous demographic data and would not enable Microsoft to identify any specific users. While gender was something the researchers could predict somewhat through analyzing Web page views, they actually found that they could not, with a high degree of accuracy, predict age from Web browsing activity using this particular algorithm.

It’s also important to know that this was simply a research project focused on demographic information, not personal identification.

Microsoft adheres to high privacy standards and is committed to giving people control over the use of their personal information. To help protect consumers' personal information, our products and services go through extensive privacy testing and review as part of our Security Development Lifecycle (SDL). We also are committed to giving people the tools and guidance they need to make informed choices about how their personal information is collected, shared and used.

We look forward to any feedback or questions you have.