AI in Policing and Discrimination: A Warning from Researchers
BBC Homepage
Skip to content
Accessibility Help
Your account
Home
News
Sport
Earth
Reel
Worklife
Travel
Culture
Future
Music
TV
Weather
Sounds
More
menu
Search BBC
AI and Policing: A Potential Pitfall
By Adrian Dredge
The growing use of AI in policing has raised concerns among researchers. Studies suggest that predictions of future crime hotspots may be based on biased data, potentially magnifying existing prejudice and creating a dangerous cycle.
Experts advocate for increased human oversight in the use of AI for policing. This includes scrutinizing data to remove any bias, as emphasized by Dr. Carolyn McKay, an associate professor at the University of Sydney Law School.
The issue lies in the fact that AI systems are trained on data that already reflects human bias. This can lead to self-fulfilling prophecies, similar to a map-maker relying on outdated maps.
Imagine AI as a map-maker using 50-year-old maps to draw a city. Now, if this map-maker is given a list of past crimes and their locations, and tasked with identifying areas prone to future crime, police patrols will naturally increase in these areas. Consequently, more crimes will be recorded in these zones, leading AI to predict even higher crime rates for them in future maps.
This problem is not new to policing, but AI may be accelerating it.
Sarah Brayne, a sociologist at the University of Texas at Austin, found in her research that the Los Angeles Police Department increasingly uses data to predict crime, with disproportionate focus on black neighborhoods. She argues that this creates a feedback loop where AI predictions contribute to the very events it aims to predict.
However, Dr. Dimitrios Kollias from Queen Mary University of London (QMUL) counters that modern AI technologies have advanced to mitigate historical bias.