Artificial intelligence has the power to transform our lives and businesses, but only if it’s used in the correct way. Artificial intelligence algorithms can help impact climate change by ensuring supply chains are efficient and streamlined and that shipping is done in the most efficient way possible. It can ensure manufacturing practices limit waste. But AI can also be used irresponsibly if we aren’t aware of the problem of implicit bias. Doomsday movies in which the robots take over and destroy humanity are there to caution us of the worst case scenario. Unfortunately many aren’t worrying about the unintended consequences before they proceed in implementing this technology in fighting crime, and the results so far have been less than encouraging.
Now law enforcement agencies are looking to use data on crime to predict future crime and prevent it from happening. The main point is to further drive down the crime rate by sending patrols to areas where crime is more likely to happen according to the data. This is based on the belief that crime leads to more crime and that crime is infectious like a virus.
Data about crimes is fed into an algorithm and the algorithm is used to predict the next place where crime is likely to occur. There are a few problems with this, however.
In one study, Amazon’s facial recognition software misidentified 28 members of Congress as known criminals. Of those 28, 38% were people of color, and people of color only made up 20% of Congress at that point in time.
When existing bias is programmed into the artificial intelligence algorithms, the output is inherently biased. As the old saying goes, garbage in, garbage out.
In Huntington Beach, California, however, a 500-pound autonomous robot is patrolling the streets using facial recognition software, license plate detection, and mobile device identifiers to report the presence of “blacklisted” persons to local police.
There are autonomous police vehicles in development, mass surveillance technology in use, and software that can pull all this data together to give police more access to information than they have ever had before in history.
Read next: Here's How Including Artificial Intelligence in a Business Can Bolster the Productivity of a Team (infographic)
How Is AI Being Applied To Crime-Fighting?
Artificial intelligence has been used to mine big data for actionable intelligence for years. UPS used it to determine how to cut down on fuel usage and make delivery trucks and their routes more efficient. It’s used in manufacturing to cut waste and in farming to maximize production.Now law enforcement agencies are looking to use data on crime to predict future crime and prevent it from happening. The main point is to further drive down the crime rate by sending patrols to areas where crime is more likely to happen according to the data. This is based on the belief that crime leads to more crime and that crime is infectious like a virus.
Data about crimes is fed into an algorithm and the algorithm is used to predict the next place where crime is likely to occur. There are a few problems with this, however.
Historic Crime Data Is Flawed
Historically, crime prevention laws have been targeted at minority groups. Because of this, existing data about crime is inherently flawed. What’s more, facial recognition algorithms do very poorly at identifying people of color. Because of these two major factors and several other minor factors, the algorithms being used to predict crime and police communities can be counted as deeply flawed.- Also read: Artificial intelligence And Smartphone Photography: How Tech Makes You Look Like A Pro (infographic)
In one study, Amazon’s facial recognition software misidentified 28 members of Congress as known criminals. Of those 28, 38% were people of color, and people of color only made up 20% of Congress at that point in time.
When existing bias is programmed into the artificial intelligence algorithms, the output is inherently biased. As the old saying goes, garbage in, garbage out.
Crime Fighting AI Is Already In Use
Unfortunately, many law enforcement agencies have thrown caution to the wind and are already using this technology to police communities across the world. ShotSpotter has been used in many communities already for years, and it works by listening for gunshots and then reporting the location of those gunshots to law enforcement. This has been met with mixed feelings in the communities in which it is used, where residents are often hesitant to call police for fear of retribution. Unfortunately, it has also been shown that police don’t respond to a large portion of these calls to look for evidence, making it little more than an unnecessary expense for many communities.In Huntington Beach, California, however, a 500-pound autonomous robot is patrolling the streets using facial recognition software, license plate detection, and mobile device identifiers to report the presence of “blacklisted” persons to local police.
There are autonomous police vehicles in development, mass surveillance technology in use, and software that can pull all this data together to give police more access to information than they have ever had before in history.
What Will The Social Impact Be?
It remains to be seen what the full social impact will be, but already there is growing mistrust between communities and law enforcement. Learn more about the hidden social impact of crime-fighting AI from the infographic below.
Read next: Here's How Including Artificial Intelligence in a Business Can Bolster the Productivity of a Team (infographic)