There is no clear evidence to suggest that the use of new predictive policing technology leads to safer communities, fairer treatment, greater transparency or increased trust in the police. This is the conclusion of a report by civil liberties and human rights organisation Liberty.
Although predictive policing is being heralded as a game-changer for fighting crime, Liberty cited a lack of transparency within predictive policing programs as a significant barrier to achieving rights compliance, robust regulation, proper oversight and greater transparency of their use.
In addition, Liberty stated that bias within the datasets used by police forces - created by gaps in the data or through discriminatory behaviour of police officers - would be amplified by the use of predictive policing programs.
The writers also warned of a chilling effect on the public from oversurveillance by law enforcement coupled with a culture of oversharing and big data. They wrote: “A culture of big data allows the state to monitor us even more closely and build up intrusive profiles from thousands of pieces of information. This chills our freedom of expression, making us feel we are being watched and forcing us to self-censor.”
The report makes five recommendations. They are:
The Gangs Matrix, based on the idea of “pre-criminality,” is a database of people that the police claim are at risk of being drawn into gangs or are at risk of gang violence. In July 2016, 78% of those included in the Gangs Matrix were black, 75% were victims of violence and 35% had never committed a crime.
The report states that the ICO carried out an investigation into the Gangs Matrix in 2018 and found that there were “multiple and serious breaches of data protection laws,” and that the Gangs Matrix did not clearly distinguish between perpetrators and victims.
The publication highlighted some worrying instances of data sharing including Durham Police’s use of the Harm Assessment Risk Tool or HART which uses machine learning to decide how likely a person is to commit an offence in the next two years. In 2018, it was revealed that the HART program was supplemented with a dataset from Experian called “Mosaic” that was produced after profiling 50 million adults in the UK.
Liberty sent freedom of information requests to 90 police across the UK. From the response, the researchers found that 14 police forces are using predictive policing programs at the moment, have used them in the past, or are taking part in relevant trials or research.
Responses from police forces that employ predictive policing programs were included in the report. These responses highlighted concerning gaps in understanding on the part of some police forces about how bias works. Durham Police’s response when asked about bias or the exacerbation of pre-existing inequalities was to say that no accuracy comparisons have been made between different demographic groups and that in the Durham Constabulary area of 90% of the population describe themselves as White British.
However, people of colour representing a small percentage of a local population does not negate the possibility that they might face discrimination from the police which could be amplified by the use of predictive policing programs.
Looking at wider issues relating to crime, the report writers said that there are other areas such as education, housing, employment and social care that should be addressed that would make a difference in preventing crime.
“While it may seem laudable to prevent crime before it even occurs, this is best achieved by addressing underlying social issues. The solution does not lie in policing by machine.”