MIT Professor Alex Pentland has a vision for data co-ops that allow members to control their personal health and financial data and benefit from its analysis in a new book, “Building The New Economy.”
Image: MIT

The new book, “Building the New Economy,” has a plan for making cities more agile and resilient: Local control of individual data. Residents of a small town or a large city could pool their data in co-ops which would analyze this data to benefit the members. Instead of a one-size-fits-all approach to healthcare needs or banking services, communities could tailor solutions to fit the local demographics and economy.

SEE: Tableau business analytics platform: A cheat sheet (free PDF download) (TechRepublic)

The geographic intensity of the coronavirus pandemic is the perfect example of the need for local solutions as some regions see high numbers of cases while other locations see little to no virus activity. Climate change is another example of a national challenge that will require different local solutions.

MIT Professor Alex Pentland; Alexander Lipton, a Connection Science Fellow at MIT; and Thomas Hardjono, director, MIT Internet Trust Consortium, edited the book. Pentland, a leader of the Sloan Initiative for Digital Economy and founder of MIT’s Connection Science program, spoke with TechRepublic about the book, which is part of the MITP Works in Progress (WiP) collection.
Pentland’s research group, MIT Connection Science, wants the same technologies that are causing social unrest to power more agile government, health, and financial systems.

SEE: Hiring Kit: Market research analyst (TechRepublic Premium)

The idea is to use distributed systems to give individuals and cities control over their own data. Right now, health insurance companies and hospitals have primary control of an individual’s health data, and banks get the most benefit from analyzing customer data. Individuals have access to the information but there’s no easy way to put it to good use.

If smaller, local organizations–like credit unions–could create a secure platform for people to manage their own data, this would shift decision-making and control to people and communities instead of national corporations.

Increasing local control of data would allow leaders and people to figure out solutions that fit the needs of their communities, instead of using a one-size-fits-all approach. Pentland used the example of the Upper Peninsula of Michigan and Boston. He grew up in a rural community but now lives in an international, urban, tech-centric city

“The rules here are totally different, and what works for the Upper Peninsula does not work here,” he said. “The idea is to handle local conditions locally and coordinate globally so cities can learn from each other but be responsible for themselves.”

The book has three sections: The Human Perspective, Resilient Systems, and Data and AI. The final chapter on Computational Law discusses how to deploy and regulate these new societal systems. The 14 chapters include: Health IT: Algorithms, Privacy, and Data; Narrow Banks and Fiat-Backed Tokens; Stablecoins, Digital Currency, and the Future of Money; and Interoperability of Distributed Systems.

In addition to explaining the big picture framework for this new approach to data ownership, the authors provide technical details about how to make this new vision into a reality.

How would a data cooperative work?

In the “Data Cooperatives” chapter, the authors use an example of drivers for Lyft and Uber pooling data among themselves to understand how drivers are paid and whether payments are the same across a single city. Pentland said this approach would help address a lack of transparency that leads to exploitation and unfairness.

“If a community aggregated their data, you could bring a lot of transparency,” he said.

A data collective created by Lyft and Uber drivers would put them in direct opposition to the ride-hailing companies, whose executives use the data to maximize company profits and set business strategy. The biggest barrier to these data cooperatives is that many businesses see this data as an asset. Making it easy for individuals and competitors to see and use this data would be a competitive disadvantage.

Pentland, who contributed to Europe’s General Data Privacy Regulation, said that the key to success with regulation is for all stakeholders to see something positive in the law.

“If everyone can’t see a win-win-win, you’ll get craziness and nothing will happen,” he said.

There are several key aspects to the notion of the data cooperative:

  • Individual members own and control their personal data
  • The data cooperative is member-owned and member-run, and has a legal obligation to act in the best interests of members
  • The co-op would provide direct benefit to members, primarily in the form of analysis and insight

The authors suggest using the MIT Open Algorithms (OPAL) approach to ensure the privacy of the member’s data held within the personal data stores. The OPAL paradigm requires that data never be moved or be copied out of its data store, and that the algorithms are instead transmitted to the data stores for execution.

The authors list these the principles as the guidelines for an open algorithms paradigm:

  • Move the algorithm to the data: Instead of “pulling” data into a centralized location for processing, it is the algorithm that must be transmitted to the data repositories endpoints and be processed there.
  • Data must never leave its repository: Data must never be exported or copied from its repository. Additional local data-loss protection could be applied, such as encryption (e.g. homomorphic encryption) to prevent backdoor theft of the data.
  • Vetted algorithms: Algorithms must be vetted to be “safe” from bias, discrimination, privacy violations and other unintended consequences.
  • Provide only safe answers: When returning results from executing one or more algorithms, return aggregate answers only as the default granularity of the response.

OPAL is a non-profit socio-technological innovation developed by a group of partners around the MIT Media Lab, Imperial College London, Orange, the World Economic Forum and Data-Pop Alliance. The goal is to unlock the potential of private sector data for public good purposes by sending the code to the data in a safe, participatory, and sustainable manner.