Police are using machine learning to help catch future criminals
Remember the movie Minority Report? The Tom Cruise, Steven Spielberg thriller proposed a terrifying future world in which police officers could burst into your home and arrest you, even though you hadn’t committed any crimes.
The story — originally imagined by science-fiction writer Philip K. Dick — envisioned a computer program that could predict future crimes and have the police pre-arrest criminals.
Wow, that sound fantastical, doesn’t it? It could never happen in Canada, or could it?
Predictive policing is a new tool that some police services are employing to help them better analyze crime patterns and be more prepared to deal with problems before they happen.
More police services around North America are implementing various forms of predictive algorithms and some are even finding success, at least in the early days of the new technology.
In Vancouver, the police department has seen a 27 per cent reduction in property crimes since it started using a system that employs machine learning to help officers know what areas of the city may be more targeted by thieves.
In Edmonton, a system was used that saved the department money: $1.60 for every $1 spent, according to a Walrus magazine feature article.
“The latest iteration of these analytics can’t ID a killer-to-be, but it can offer insight into what areas are potential sites for crime by drawing on information in everything from historical records to live social-media posts,” wrote John Lorinc in The Walrus.
The city targeted a neighbourhood around Boyle Street, which had been plagued by drug dealers and petty thieves.
But as Tim Wu wrote in The Globe and Mail, China is currently using a similar technology to track and potentially apprehended Muslim Uyghurs in the country’s remote western region.
“Those who amass enough suspicion points on an algorithmic scale become targets for an involuntary ‘re-education’ program – in other words, arrest and detention based on future dangerousness.”
This nightmare scenario might be a future predictor of what governments have the power to do.
How much of a stretch would it be for a Canadian politician, using populist rhetoric and overzealous predictive policing programs, to demonize and target for harassment certain populations who disagree with government orthodoxy?
Sure, it sounds far-fetched, but we are Canadians should not let our rights to privacy be frittered away so quickly.
“But if anything, given how wired the population is, the data available on North Americans is actually more detailed and complete. It just needs someone – law enforcement – to bring it all together,” wrote Wu, a Columbia University professor.
In Ottawa, the police service began using a tool in 2016 that may be targeting protesters, which is a perfectly legal thing to do in a western democracy, via social media.
"We've been talking to activists who've experienced surveillance and [they say] it makes them think twice about protesting," said Brenda McPhail, privacy director for the Canadian Civil Liberties Association.
South of the border, the U.S. government is funding some research by predictive-policing pioneer PredPol to quantify gang memberships, which come with greater legal punishment in California, for example.
Critics say this non-human way of identifying potential gang members completely wipes out the community-policing model that is employed by most progressive police departments.
“This new line of research suggests that [PredPol founder Jeff] Brantingham has not taken critiques of his research methodology to heart and is pressing forward with a project that is founded on incomplete data, dubious methods, and a premise that, if applied in the field, could result in more people of color behind bars,” wrote Ali Winston and Ingrid Burrington in The Verge.