Imagine the following: a group of teenagers drive up to a bank in a white van, having the intention to rob it. Yet somehow, there are already multiple police cars parked in front. Now knowing the futility of their effort, they drive away. What could have informed those officers to be there? How did they know that crime was going to happen?

What you just imagined was what very well could be the result of predictive policing, a new and developing technology designed to discourage crimes before they happen. The algorithm takes pre existing data on crime and analyzes it via three categories; time of crime, type of crime, and location of crime. Using that data, it creates, a digital map of locations where crime is likely to happen. In one study it was found that the software was only 4.7% accurate in predicting crime, however it was still higher than the 2.1% of human analysts.

Despite it’s perceived advantages, this type of policing can be seen as an effort to spy on citizens, or as an excuse to target low-income neighborhoods. And this may be correct, as the software has been seen to target low-income neighborhoods more. While this isn’t intentional bias, most officers claim, many groups perceive it as the result of over-policing areas in the past.

It is very easy to understand why technologies like these stand out, yet this predictive policing has a critical flaw.  It cannot account for the bizarre, unusual, and extreme incidents that occur more than one might guess. Even far from the large attacks that might immediately come to someone’s mind, somehow more unexpected crimes can occur. One such crime was bizarre, destructive, and was made in someone’s garage.

Enter what is now nicknamed the “Killdozer,” a homemade tank that caused–despite no injuries or casualties–around seven million dollars in damage in Granby, Colorado. The tank was the brainchild of Marvin Heemeyer, a man who described himself as “a reasonable man driven to do unreasonable things.” Disregarding motives, the tank was constructed from a Komatsu D355A bulldozer; homemade armor was fitted over the dozer itself. The armor made out of two thick sheets of steel with concrete in between, was equipped with three gun ports, and exterior cameras covered with bulletproof plastic and even having air nozzles to blow dust off of them. Inside were three monitors for Heemeyer to see where he was driving, and to top it all off, an air conditioner.

Any and all attempts to stop this humongous machine were futile. Armor-piercing ammunition even failed to penetrate the cabin. An attempt was made to throw a flashbang grenade into the exhaust, but was met with no reaction. Anti-tank missiles were deemed to be unusable due to causing more damage than Heemeyer could have ever caused. The machine was only ever stopped after it got stuck and the engine overheated. In the end it took 12 hours with an oxyacetylene torch to break through the armor, explosives having failed to do so much as scratch it.

Yet another bizarrely technical case is that of the Harvey’s wagon wheel casino bombing. The bomb itself, constructed out of stolen dynamite, contained several mechanisms to prevent any sort of tampering, not even allowing the bomb to be so much as moved. These mechanisms proved to be successful, seeing as the bomb detonated in an attempted disarming.

In the case of crimes like these creating copycats, would it benefit the police forces involved to have technologies made to beat these extremes? Maybe a mobile drill vehicle for the Killdozer? Or a complicated power tool that could bypass the security of the wheel bomb? Or would it be argued that these technologies would give the police too much power? I think most would rather be grateful for the additional security any new technology would give me.