Data collaboratives try to make data collection work for public good

When you take an Uber ride, information about your trip may be stored and shared with governments. But it’s not fed into a controversial mass surveillance system or security apparatus; instead, the data is anonymized and analyzed by cities, to help with problems like traffic flow and congestion.

It’s called “Uber Movement,” and is like hundreds of similar initiatives around the world. These programs, called data collaboratives, take the mass data collection of tech companies and try to squeeze some public good out of them.

An interesting piece of analysis this week looked at how data exchanges could enhance predictive capacity. “By harnessing the power of data collaboratives, governments can develop smarter policies that improve people’s lives,” wrote Stefaan G. Verhulst, co-founder of NYU’s GovLab, which tracks data collaboratives.

For instance, the Uber Movement program has helped Ohio officials design better traffic systems to reduce road deaths.

However, amidst the ongoing “techlash,” data sharing is increasingly seen as a dangerous and unaccountable practice. It’s precisely this kind of data collection that has let Silicon Valley monetize private information and commoditize the public sphere. Even the public-good rhetoric of data collaboratives can be misused, as we’ve reported in the case of the global, highly uneven rise of “smart cities.” In that case, improving urban planning has justified the deployment of surveillance states.

Verhulst told me that public concerns about mass data collection have made data collaboratives more difficult to set up.

“Cambridge Analytica, along with other data breaches, have generated an overdue privacy awakening among decision makers and the public at large,” Verhulst wrote in an email. “At the same time, the collateral damage could be huge and damaging to legitimate scientific initiatives and the generation of important societal insights that could improve people’s lives and governance.”

He also said that some projects have already faced challenges.

“Some data holders have become more reluctant to respond to legitimate research and data sharing requests; while some researchers have become more reluctant to become associated with data initiatives that leverage private data,” he said, adding: “As a society, we need to become more sophisticated about the real value and the real challenges of data sharing and not be one-sided.”

Still, there are legitimate political and ethical concerns about the sharing of private information with third parties, even when done in the name of better governance. Verhulst suggests that data collaboratives, which are inherently built to serve the public good, can serve as models for how to build transparency and accountability into surveillance systems that usually have neither.

“There are ethical and privacy concerns across the data life cycle that could lead to discrimination, harmful data disclosure, increased inequality,” he said. “As such we have called for a new profession to be established: a Chief Data Steward that can consider the value of private data for society while limiting the risks and preventing misuse.”