Open government data – more critical than ever.


Today, in Washington, D.C., Microsoft was pleased to participate in an event hosted by  the Business Software Alliance, focused on Data Innovation Policy:  Enabling Access and Promoting Use. We were honored to have U.S. Rep. Derek Kilmer (D-WA) provide introductory remarks. Rep. Kilmer is a strong advocate for open data, having served as sponsor of the OPEN Government Data Act, which was signed into law in January 2019. As Rep. Kilmer noted, Congress and the administration have recognized that the availability of useful government data is essential for the U.S. to lead a digital economy powered by AI and data analytics. The OPEN Government Data Act’s mandate – to encourage every federal agency to publish information as open data – is fundamental to achieving this goal. This mandate is ambitious and presents a range of policy, structural and technical challenges. Multiple agencies need to develop and implement effective approaches to identify, maintain and publish relevant data inventories, in a standardized, machine-readable format. Important progress has been made toward these ambitions.

And yet, there is more that can be done to achieve this vision. One idea I mentioned at the event is the idea of creating a Federal Chief Data Officer role to help spearhead the goals of the OPEN Government Data Act. The creation of such a role would help agencies coordinate and prioritize the work to unlock high value government data.

At the event today, we heard about many compelling examples where open government data has been used to advance research in important areas. For example, our speaker from the Fred Hutchinson Cancer Research Center spoke about how government data was being used by a scientist to help look at new ways to identify and treat endometriosis. We also heard from Rep. Kilmer about how environmental data was being used to help forecast weather and transportation trends.

We can also make progress by aligning on standardized terms and agreements to make it easier for stakeholders to use and share open data. In July, we shared three draft data use agreements for community input. The agreements govern the sharing of data, particularly in the context of training AI models. There is more work to be done in this space, and groups like New York University’s  GovLab are also making important strides in driving toward common terms.

Of course, data sharing and collaboration must take privacy interests into account. To help ensure robust privacy protections, we should also work to help governments collaborate with privacy-protecting technologies such as differential privacy, which is being used by the U.S. Census Bureau. Such technologies, when used appropriately, can provide important privacy safeguards, while still enabling data to be analyzed for the benefit of all.

Alongside stakeholders from industry, the nonprofit sector and academia, we are ready to work with government to take a deeper look at how we can unlock data for AI innovation and economic opportunity. Today’s discussion was one step along this path, and we look forward to more to come.

Image credit: