Big data is transforming policing. Predictive technologies can forecast crimes yet to be committed. Data on the location of crime is shaping when and where police patrol. And intelligence systems can highlight offenders who are responsible for a disproportionate number of offences.

The potential benefits are immense. In Colombia, better use of data helped police slash the country’s murder rate by 82% in a quarter century. But the legal, political and economic problems posed threaten the balance of power between citizen and state the world over.

Andrew Guthrie Ferguson’s The Rise of Big Data Policing: Surveillance, Race, and the Future of Law Enforcement carves a space between the uncritical optimism of many law enforcement advocates and the trenchant criticism of their opponents. The data revolution, he argues, is underway: the only way to maximise its benefits and minimise its costs is through a constant process of public accountability.

• For more like this, see our violence prevention newsfeed.

Apolitical spoke to Guthrie on the promises and pitfalls of big data policing.

How is big data changing policing?

Data helps police decide where to go, who to target, when to patrol, and how they investigate crimes. It’s increasing their surveillance powers, and helping them to see different patterns of crime emerging in society. But it’s also changing the ways police and citizens relate in ways we’re only just starting to recognise and evaluate.

In the book, you argue that data can’t be spoken of as one as one homogenous mass of information. You propose the idea of “black data” as one important category. What does it mean?

“Black data” is a term I use to describe data affected by three interrelated issues.

First is the issue of transparency. These are black box systems — secret algorithms in many cases — which are difficult to understand and yet are changing the way we are policed.

Second is the issue of race. In the US, the data we use is racially encoded: the data reaching these systems is coming from police officers and in many cities that’s very much tainted by racial biases.

The final reason for calling it “black” is because it’s distorting. Our laws and frameworks were developed in the small data age and we’re now facing a big data future. We haven’t yet developed an understanding of what this data might do to society, and it can distort our vision.

The idea of black data tries to bring together interrelated issues of transparency, racial bias, and legal distortion to give us all pause before we embrace this technology. I argue that pretty much all policing data at the moment is black data.

But black data isn’t the only kind of data you discuss. You propose the ideas of “blue data” and “bright data” as more promising alternatives.

As we’ve built this immense architecture to surveil citizens, we’ve also accidentally built an even better architecture to surveil police officers and police conduct. If you think about a modern police conduct, they’re driving in a GPS-connected car, we can see what their tactics are and where they’re going. They’re watched on body cams, and we can see what they’re saying to people. This is what I call “blue data” — which can and should be used to force the police to be accountable.

Ironically, there’s a huge pushback against “blue data” from within law enforcement: just as citizens don’t want to be surveilled, police don’t want to be surveilled either. We’ve created the technology that could be used to make police officers more accountable, but we’re hardly using it because of the resistance.

“Bright data” refers to the clarity that good data can provide in devising the most appropriate solution to a given problem. Crucially, it can suggest modes of response outside of policing. If the data shows you that crime is concentrating in a certain area, “bright data” could suggest that you fix up the neighbourhood rather than just putting a police car there. If you know an individual is at risk, maybe you send a social worker rather than an officer.

How could a focus on this “bright data” change predictive policing?

I once joked that the problem of predictive policing isn’t the predictive part — we’re very good at that now — the problem is the policing. We don’t need to respond to all problems with the police, and data should help us understand that and factor that into decision-making.

Another concern you raise is who gets to own the data.

One of the untold stories of the rise of big data policing is that it has been very much driven by startups pitching local police chiefs. Those officials often haven’t been asking the tough questions they need to about who owns the data and what it will be used for.

In a world where police are collecting more and more information, that data can be of immense value commercially. We haven’t thought through proper contractual arrangements regarding how we keep public data in the public sphere and not monetise it.

One of the newest frontiers is the police body cam space. In the US, Axon (formerly known as Tazer) have just bought out their rivals and are reaching a quasi-monopoly state. They have immense amounts of data from all these body cams on officers across America, which amounts to involuntarily obtained public data about public cities.

There are parallels to the whole Facebook and Cambridge Analytica mess. The outrage wasn’t just that Facebook had this data, but that it ended up in the hands of unknown third parties. That’s exactly what’s happening in the government space.

I’m not worried that body cams are providing information to the police to help them do their jobs, I’m worried about that data not being controlled, letting private companies monetise it however they want, leaving the citizen with little stake in how their data is used and infringing the privacy protections that exist.

What’s the solution?

The irony of all this fancy new technology is that its future is going to come down to that government lawyer, buried in a bureaucracy somewhere, who is asked to draw up contracts between cities and tech companies. Those contracts have never been sexy or exciting. But they’re crucial.

If city governments began to realise that they have something valuable, they could simply stop contracting with entities that seem like they’re offering something for free, when really they get to keep our data and monetise it later.

But most crucial of all is the moment of public accountability. In the current political environment in the United States, the only hope is in local government. In Seattle, Oakland, Boston, and Berkeley, we’ve seen local communities demand public oversight of new policing technologies. That public moment of accountability provides transparency that still harnesses the benefits of technology, but uses in ways that are open and transparent.

You need to have conversations right from the off: how are we using this technology? When are we using it? Are there certain places we don’t want surveillance coverage? What about outside medical clinics? Churches? Mosques? Civil liberties charities? How will this impact on associational freedoms, or freedom of protest?

This can have a happy ending: we just need some transparency so that public figures with public accountability can make these calls on what they think is a fair balance of privacy and security, public welfare and public safety.

(Picture credit: Flickr/Evgeniy Isaev)

Edward Siddons
edward.siddons@apolitical.co