Update on the Collaboration between Harvard’s Institute for Quantitative Social Science and Microsoft to Develop Open Source Differential Privacy Software

June 24, 2020
Open DP logo with starburst icon and name

In September 2019, Harvard’s Institute for Quantitative Social Science announced a large-scale collaboration with Microsoft to develop open source tools for differential privacy. The promise of differential privacy is that researchers can discover patterns in data while simultaneously providing mathematical privacy guarantees for survey respondents and others whose information may be represented in the data. We are excited by our progress and write today to report on several lines of substantial progress.

First, we are announcing Microsoft’s grant of a royalty-free license for its differential privacy patents to the world through OpenDP. OpenDP—which is being incubated by Harvard’s Privacy Tools and Privacy Insights projects (at SEAS and IQSS), with generous support from the Sloan Foundation—is an effort to engage a community of collaborators across academia, industry, and government to create trustworthy open source software, free to all. This is a significant contribution to the world and our work, not the least of which is because the Microsoft Research Silicon Valley lab was the birthplace of differential privacy a decade and a half ago.

Second, we have released, with Microsoft, OpenDP’s first end-to-end operational differential privacy system. This software is usable for certain low-risk applications with a trusted analyst now and we expect it to grow and improve over time as we see how it operates in the field. It will be adapted to integrate with the OpenDP Commons that we collectively build over the coming months. OpenDP has a long road ahead of it, but this is an important step.

And finally, along with these developments, we have joined our workstreams so that the IQSS-Microsoft collaboration and our OpenDP project are now merging into one powerful force, the broader OpenDP community. This community already includes hundreds of participants who attended our recent OpenDP Conference.  We continue to build out its committees and foster stronger ties with researchers across the landscape of industry, academia, and government.

We look forward to building out this industry-academic collaboration, modeling it in other projects we develop, and seeing what we do together next.

Gary King, Director Institute for Quantitative Social Science, Harvard University