Hitachi Claims Predictive Analytics Can Be Used to Predict Crime
The world depicted in the Minority Report movie — now its own scripted television series — might not be as far-fetched an idea as you'd think. Hitachi is developing new technology that they claim can predict crime before it even happens.
Hitachi is currently perfecting a new program that analyzes an almost infinite amount of information in order to pinpoint where and when crime will strike. According to Quartz, Hitachi Visualization Predictive Crime Analytics (PCA) uses information collected from gunshot sensors, license plate scanners, weather forecasts, public transportation routes, and countless other sources to assess the risk of criminal activity in a given location.
The system is designed to detect eminent threats in real-time, effectively eliminating guesswork on the part of law enforcement. It’s also just the latest in a string of innovations involving predictive analytics, the use of computer algorithms to pull meaningful insights from massive troves of otherwise unintelligible data.
Interestingly, social media is set to play a prominent role in the program's crime prediction algorithm. According to Digital Trends, PCA is able to decipher messages sent over social media platforms that are written in codes specific to different gangs. What’s more, the machine is capable of learning, and takes new statistics and variables into account as time goes by.
Although the idea of a super-intelligent computer does sound a bit like the set-up for another blockbuster doomsday film, the truth is that PCA is very similar to technology we’re already well familiar with.
Predictive analytics also help Netflix predict the movies you would enjoy, and determines what news and social stories end up in your Facebook feed. The algorithm is prone to the same biases as the person who created it, which is to say that charging a computer — rather than a human — with the task of crime monitoring does little to increase the risk of bad policing.
In fact, there are a number of obvious upsides to implementing a system like PCA. Police officers, who usually rely on past experiences to predict crime hotspots, are freed up to spend time preventing or preparing for likely offenses.
Moreover, because the program predicts what kinds of crimes are most likely to occur, officers will be able to devote time to developing more specific plans of action oriented toward particular crimes, rather than having to be prepared for anything and everything at a given moment.
The public, however, has raised some other, very plausible concerns about the use of PCA to catch criminals. Profiling and stereotyping is one recurring fear: if the system relies largely on where crime is "most likely" to strike and who most "fits the profile" of a criminal, people and communities could be unjustly targeted by officers based solely on location or identity.
Hitachi reacts to such concerns with claims that PCA will actually eliminate profiling by providing officials with factual data (rather than stereotypes) to draw from when making arrests. This raises another concern — how accurate can the program actually be? Is it possible that it could make largely faulty predictions that will detract from dealing with real crimes? Only time will tell — Hitachi is currently working on trial runs before the system goes public.
No matter what results the test period yields, it's easy to see how programs like PCA can be helpful in a variety of fields. In almost any sector, predictive analytics can drastically cut the need for manual calculations.
And while the success of predictive analytics in the world of law enforcement has yet to be seen, it has vastly improved the capabilities of those in IT. As the computer systems that enterprises rely on become more complicated, the data that IT professionals must track has grown impossible to interpret.
Automated predictive analytics, like those available through TeamQuest’s AutoPredict software, make it easy to tell where disasters will occur. The promise to prevent crime through data is exciting, but when it comes to server crashes, that promise is already a reality.
(Main image credit: Bill Dickinson/flickr)