Police algorithms should be regulated, says think tank

Rising volumes of digital forensic evidence, public demands to engage with the police online, and reduced budgets need to be handled with more “imaginative” use of data, along with regulation, according to research by independent think tank the Police Foundation.

The report, Data-driven policing and public value, outlined various recommendations around integrating algorithmic decision support tools into policing practice, and emphasised the need to introduce of new rules to govern the use of such technologies in policing and the criminal justice system.

Acknowledging the concerns the public has about use of data, the report states there is a need to insist that software used for data-driven decision making is transparent and available to independent checks.

“Crucial decisions in the criminal justice system that affect lives cannot be left to unseen and unchallengeable ‘black box’ processes,” the report said.

The recommendations were published as it emerged that the Metropolitan Police Service is secretly developing a new criminal database with similarities to the controversial Gangs Matrix, which has been raising concerns among data protection and racial equality activists.

The Police Foundation report also recommends banning any systems that are driving decisions within the criminal justice that cannot be “deconstructed to show how predictions of possible future behaviour have been arrived at in individual cases”.

Additionally, it advises the enforcement of a requirement whereby decisions driven by algorithms are made in a simple language so that defendants can understand and challenge them.

The research also questioned the ethics and public anxiety around algorithm use in decisions that “can have profound implications for both procedural fairness and individual human lives”.

“Clear regulations around these issues would help to ensure that citizens and their legal teams can understand and challenge new processes and tools and therefore hold the police accountable for their use,” the report said.

Additionally, the paper listed a range of challenges such as data misuse by police forces, implications for personal privacy and building predictive models on the basis of inevitably biased and inaccurate data.

Creating privacy and ethics commissions into the governance structures of police forces to address privacy concerns around surveillance technologies is a related recommendation set out in the report.

The think tank also suggested that UK policing adopts a common set of data standards and entry codes to be used across the country.

“The Police ICT company should be given the role of developing one and its subsequent use should be mandated across all police forces,” the report said.

“A common set of access protocols across all police forces are also needed so officers can be sure that other forces are not only capturing the same data, in the same way and in the same format, but that officers of the same rank and role are engaging with that data too,” it added.

Purchasing “closed technology”, so any tool that can’t be simply implemented and is not interoperable, should also be banned in UK police forces, the report said.

Police officer training in data-driven activities is another recommendation listed in the report, as well as educating citizens about the complexities surrounding data-driven policing, so they know what is and isn’t acceptable in terms of big data applications within police practice.