Here’s how the UK public sector is using artificial intelligence

AI public sector

The use of artificial intelligence (AI) could add a whopping £654bn to the UK economy, according to a report published by Accenture.

Perhaps this is why the government announced late last week that it would pump approximately £20m into developing and researching AI and robotics as part of its long-awaited Digital Strategy.

The technology, which could potentially render almost half a million public sector workers unemployed over the next 15 years, has been credited with helping organisations save time and money – both attractive propositions for governments across the globe.

With this in mind, here are four examples of how the UK public sector has delved into the world of AI in recent times.

Serious Fraud Office

The Serious Fraud Office used AI for the first time in its landmark investigation of Rolls Royce, which saw the car manufacturer found guilty of bribery and corruption in January this year.

The robot, developed by UK AI startup RAVN, helped the government’s law enforcement arm sift, index and summarise 30 million documents relating to the investigation.

According to the Financial Times, the robot was intelligent enough to build on its understanding of the case and identify relevant material.

In fact, the company’s chief executive David Lumsden said the robot contributed to cracking the case “exponentially quicker” than if it had been investigated manually.

Local government

Enfield London Borough Council made an interesting move in 2016 – it purchased Amelia, a robot capable of making decisions and tracking customer emotions.

Developed by US-based labour tech company IPsoft, the robot will be used by the council to help improve service delivery by directing residents to correct information and helping to authenticate licenses.

Amelia, which is thought to be 60% cheaper than employing a real human, has been designed to emulate the way in which people communicate using natural language.

Speaking at the time, James Rolfe of Enfield council told the Evening Standard that there were “no plans” to get rid of any of their 50 call centre workers, adding a customer “shouldn’t see they are interacting with a digital agent, it should be a seamless experience”.


London-based Moorfields Eye Hospital partnered with Google’s DeepMind last year in a bid to use AI to help in the diagnosis of two eye diseases: diabetic retinopathy and age-related macular degeneration.

The purpose of the partnership was to determine whether or not AI can be used to reduce the time it takes to analyse complex scans, which would in turn result in faster diagnosis and treatment for patients.

Although the partnership captivated much of the UK’s media, it soon proved controversial as some privacy campaigners criticised the fact that patients were not being asked if they wanted their data handed over to the tech company.

When Google and Moorfields caught wind of this, they explained that as the data was anonymised, they were under no obligation to ask patients’ permission.

Earlier this year, some NHS’ A&E departments in North London also trialed an AI system developed by medical startup firm, Babylon Health.

As a result of the partnership, some 1.2 million patients have been able to input their symptoms into Babylon’s AI system, which then responds with questions in order to determine the seriousness of the condition or injury.

The point of the exercise is to help release the burden on the traditional 111 hotline – the NHS’ non-emmergency


Back in 2004 the Home Office’ scientific development branch and the Metropolitan Police provided initial funding to help develop radar scanners that use AI to detect hidden weapons carried by people.

Although not yet in use, the team that developed the device at Manchester Metropolitan University, said in 2013 that it may eventually used in airports and other public places.

The device uses AI to differentiate between non-threatening items like household keys and mobile phones and items that could be threatening.

Real-life testing of the devices is believed to have started in 2014.