PoliceTech report looks at how startups can work with UK police

Think local

“Here’s the news: the police system and the taxpayers have been utterly screwed over in the past decade through inefficient procurement and a seller’s paradise, so it’s time for change,” said policing minister Nick Hurd at the report’s launch event.

“The challenge is how to create an ecosystem that connects the innovators in the police system to the people who know what the problems are.”

One of the report’s authors, Blair Gibbs, added: “My view is that none of those barriers that still exist are a reason why a police force can’t form a partnership with an interesting SME. I think in some ways we’re asking people to take a different perspective.

“There will always be a place for national systems [like the Home Office at the centre], but when you look at places like North America, it’s the other end of the extreme – they’ve got a very hyper-localised system,” he said.

“One advantage of a very fragmented system is of course for the SME market – you can find many more buyers at a lower level that you can at least prove your product with.

“That being said, policing in the UK is interesting, because even the smaller county forces are big enough that the SME that wins some important project with that particular force can scale over time.”

Gibbs concluded that SMEs should focus on building relationships with local forces and Police and Crime Commissioners where they will be able to prove their product much more easily.

Further challenges

The report does warn, however, that shiny new tech is not a silver bullet. “A single product or service is never going to solve all your problems, and policing is too complex a discipline to be radically improved overnight by any one piece of technology.”

The use of algorithmic data, for example, presents a number of practical, legal and ethical challenges for any industry, but particularly in policing, where public trust is essential for maintaining legitimacy.

“Machine learning tools are very intelligent in some senses and very stupid in others,” said Jaime Grace, a senior lecturer in law at Sheffield Hallam University.

“Because they are just bits of software, they don’t understand the human biases that go into these big sets of data that they work from, so the only way you can tackle risk of retrenching discrimination in the use of these tools is to deal with the police officers that record the info in the first place.”

Grace said that any tools like this should be built in such a way that they can monitor and take into account whether or not the information used has been reviewed or cleansed.

“Unfortunately the courts are very reluctant to put pressure on police to cleanse intelligence databases. The last time this was challenged in the supreme court was 2015, and the police actually ended up with an even greater leeway than they thought they had to retain what they call ‘soft intelligence’ about individuals who might be linked to crime – even where those individuals are not guilty of any crimes themselves.”

Accountability in the public-private partnerships

Police forces should also be aware of how a company could be perceived when entering into partnerships, as any moral or ethical issues that arise could potentially undermine public trust – particularly in already marginalised communities.

For example, while they are only mentioned briefly in the report, predictive policing company Palantir provide and run software for the NSA and ICE on behalf of the US government.

Many police forces in the UK already have contracts with companies that also sell technology to countries with questionable human rights records.

Cellebrite, for example, provide police forces (including West Mercia Police, Avon and Somerset Constabulary, and City of London Police) with a range of tools to help them extract data from mobile phones and other personal devices, which they have also sold to Turkey, the United Arab Emirates and Russia, according to 900GB of data acquired by a hacker last year.

“As long as police forces are open and honest about how they have used a private company to build a tool for them, and we know in a transparent way how that private company is building pools of data into the databases or sets of big data that the tools use, then you can start to get that cycle of accountability going,” said Grace.