We aim to work with industry to develop privacy-preserving tools and research at MIT that will help shape the future of data governance.
The confluence of powerful data analytics, artificial intelligence, and global information platforms together with changing social attitudes about personal data are transforming both the policy and technology landscape. To address these global privacy challenges, the MIT Internet Policy Research Initiative (IPRI) and the Computer Science and Artificial Intelligence Laboratory (CSAIL) are launching the MIT Future of Data, Trust, and Privacy, a lab initiative with the goal of bringing state-of-the-art MIT computer science research together with world-leading public policy expertise and engagement. The initiative is an opportunity for industry partners to work with MIT researchers on new policy-informed, technical approaches to today’s privacy challenges; to understand the implications of new laws such as the GDPR; and to lead a global dialogue with policymakers, civil society, and industry leaders as we shape the future of privacy and data governance. We are inviting a small group of organizations to help us build this research initiative.
The MIT Future of Data, Trust, and Privacy initiative will lead technical research on privacy-enhancing data systems and analytic techniques, create a policy dialogue with extensive public engagement, and develop new educational opportunities related to data governance technology and public policy.
- Database Systems: Develop new data management architectures to provide enterprises with purpose management, provable delete and automated audit accountability tools for managing personal data according to legal rules and institutional commitments: GDPR, along with other current and proposed privacy laws, pose data governance challenges that cannot be met with existing database architectures. By combining the best database systems expertise with policy awareness, we will develop models for new systems that can help govern data at scale and enable engagement with policymakers about the most effective approaches.
- Applied Cryptography: Deepen the application of privacy-preserving cryptographic techniques to real-world policy challenges associated with handling personal data: By combining scientific insights into cryptography, we can build usable systems and associated policy frameworks for working with de-identified data. The world’s leading privacy laws look to private computation and other cryptographically-powered data handling techniques to enable uses of personal data while limiting privacy risk. We will bring together the cryptographers and public policy experts to expand the technical options available and contribute to the public policy dialogue on this question.
- AI and Machine Learning: Develop privacy preserving, trustworthy machine learning (ML)systems meeting global legal requirements and providing explanation and bias assessment: The sustainable and trustworthy growth of ML systems depends on both technical advances in how personal data is obtained and handled in the ML pipeline, as well as public policy dialogue to agree on norms that meet society’s expectations for both privacy and advancing human knowledge.
- Data Portability and New Information Architectures: Design new protocols for managing personal data flow across APIs to enable support for data portability requirements while maintaining usage limits and accountability: Data portability has clear benefits for competition and significant impact on privacy, perhaps positive. We will explore new technical approaches to greater individual control over data with an eye to the underlying privacy risks and benefits.
Policymakers from around the world seek out MIT expertise on issues such as encryption, privacy, and AI Policy and are an important source of inspiration for researchers. The initiative will provide a strategically managed forum for dialogue amongst MIT researchers, policymakers, industry consortium members, and civil society partners.
For example, MIT’s 2019 AI Policy Congress, a large-scale, international conference that brought together distinguished faculty, leading members of industry, and a high-level expert group of international AI policy experts from the OECD’s 36 member countries. The conference included a workshop designed to demystify machine learning and sparked a discussion on the challenges to developing effective public policy in the area. This conference and IPRI’s engagement with the OECD was a key factor in the eventual shape of OECD’s Recommendation AI Policy Principles.
The initiative will contribute to expanding academic offerings on privacy topics, and will create a professional education series delivered online and in-person that is designed for privacy professionals in industry, civil society, and government.
At MIT, we are assembling a group of leading privacy, data, and AI researchers to work on these challenging issues using a cross disciplinary approach.