Maximizing Impact through the Data Privacy Collaborative

Analyzing sensitive data without revealing private information is crucial for extracting valuable insights on workplace inequalities, transportation policies,  health care outcomes, and more.

Researchers are harnessing the power of big data while maintaining privacy through the new Data Privacy Collaborative at the Hariri Institute to maximize their societal impact. The Collaborative, led by Mayank Varia, aims to foster cooperation among the Institute, industry, other nonprofits, and government entities to further the development of open-source platforms and to deploy at-scale software pilots that demonstrate and promote the responsible use of private data assets in real-world applications.

We asked Varia about data privacy and his goals for the Collaborative in a recent Q&A:

1. What is data privacy?

Data privacy is a complex, multi-disciplinary question that spans computer science, ethics, law, and more. Roughly speaking though, I think of data privacy as being respectful of how people would want their personal information to be used. Data privacy includes only using data for the purpose of learning insights that everyone wants, only releasing data that is de-identified and cannot be used to reconstruct personal information, not collecting more information than required to perform the desired analysis, and protecting data while in use in order to reduce the chance of data breaches.

2. What is multi-party computation? How does it relate to data privacy?

Secure multi-party computation is a tool from cryptography that allows for many people to analyze their joint data without the need to share it with anyone. It is a tool that can be used to release aggregate statistics that are of common interest without revealing the underlying, sensitive data that were used to calculate them. For example, we at Boston University have used secure multi-party computation to determine the overall gender and racial wage gap in the greater Boston area without the need to share employers’ payroll data  and to measure the rate of engagement with minority-owned businesses.

3. What makes the Data Privacy Collaborative unique?

Our goal with the Data Privacy Collaborative is to go from theory to practice and back again. On the one hand, we want to understand the business needs of our industry partners and identify opportunities to transition technology to them, building upon our work over the past 5+ years in building open-source products for data privacy. More importantly though, we want to understand which challenges from our industry partners cannot currently be solved because they require new scientific advances or cross-disciplinary collaborations, which can therefore lead to new research projects.

4. What are your goals for the Data Privacy Collaborative?

I am excited about the diversity of opinions that we will learn from in the Data Privacy Collaborative. Our founding members have extensive experience in sectors of the economy like healthcare and technology, and they are intimately familiar with data privacy issues that their organizations face. The Collaborative is a great opportunity for us at BU to learn from their experience and to ensure that our research has maximal impact.

5. What impact could the Data Privacy Collaborative’s work have on society or the world?

The Data Privacy Collaborative will allow us at BU to design, develop, and maintain software products that use data privacy tools like secure multi-party computation, fully homomorphic encryption, differential privacy, and more. It will also allow us to identify concrete use-cases in which to deploy these technologies with current or future members of the Collaborative.

Interested in learning more about the transformational science happening at the Hariri Institute? Sign up for our newsletter here.