Child safeguarding transformed for police with ‘game-changing’ Artificial Intelligence

Artificial intelligence (AI) software which can identify vulnerable children months before they might be at risk is now in use by Lincolnshire Police.
Lincolnshire Police are using a new Artificial Intelligence program to assist in identifying vulnerable children at risk.Lincolnshire Police are using a new Artificial Intelligence program to assist in identifying vulnerable children at risk.
Lincolnshire Police are using a new Artificial Intelligence program to assist in identifying vulnerable children at risk.

This cutting-edge approach to child safeguarding is thanks to new ethically designed technology to identify Lincolnshire’s most vulnerable children – and the county’s police force claims it means they will be better protected.

In testing, the software, called CESIUM, used live data to identify 16 vulnerable children who needed intervention from specialist services up to six months before the existing process, which is reliant on people examining and assessing information. Further analysis proactively identified three vulnerable children for pre-screening risk assessment who had previously been concealed in the data.

The software will mean staff have more time to dedicate to other priority areas of work because less is spent on administration; the app allows a reduction from five people researching, gathering and analysing data over five days, to one person analysing all near-real-time data within 20 minutes.

Following the positive results, CESIUM, which has been developed for Lincolnshire Police in partnership with Trilateral Research, has now been adopted and will be used to help officers and staff in their decision-making around child safeguarding. In response to calls that some children might get missed, the force insists it will not replace human thought and analysis, rather it helps support and inform decision-making, improving the day-to-day workload of police officers across the force.

Detective Chief Superintendent Jon McAdam, who has worked with program developer Trilateral on CESIUM, said: “Using the technological advancements of Artificial Intelligence and Machine Learning, we have been able to work together to take what began as an idea to the actual delivery of a cutting-edge safeguarding platform.

“The development of CESIUM has only been possible through the commitment and determination of all those involved, and I am proud to be able to have finally seen this not only be successfully developed, but to bring this into the operating environment within Lincolnshire Police.

“This type of technical solution will never replace people. Instead, it helps speed up and strengthen our processes and gives our staff more information on which to base their decisions. We are now able to be more proactive in protecting vulnerable children from harm and ensuring professionals are informed of risk concerns. This has been a significant journey, where we have been able to bring the experience of people who work together in safeguarding with the technical expertise of data scientists, data ethics, and GDPR specialists, along with the wider team within Trilateral Research.”

Conversations about the system began following the release of 2018 statutory guidance on inter-agency child safeguarding when experts in Trilateral recognised a need for a more rigorous approach to the information exchange and identification of risk within local organisations working in partnership on safeguarding children. Using public funding that is designed to drive development of game changing innovations, Trilateral worked with Lincolnshire Police to co-design this AI solution.

Trilateral’s multi-disciplinary team, which includes people with expertise in social science, data science, ethics, law, child safeguarding, data protection, AI and software engineers, have developed CESIUM over the past five years.

Dr Hayley Watson, Director of Sociotech Innovation at Trilateral Research, said: “We take a rigorous sociotechnical approach to the design of ethical AI that considers both the technical and social considerations of its use in complex areas like policing. This builds trust in the use and adoption of AI, and confidence in the insights and interventions that can be achieved.”

Chris Haward, Chief Constable, Lincolnshire Police, added: “The software enables us to identify vulnerable children months before the existing process and our staff had been undertaking this manually, so the software can now support them in their decision-making. The ability to detect more quickly and accurately where we need to put prevention measures in place is game-changing.”