Opinion: Three ways AI can help tackle the prison crisis

We need a long-term solution to prison overcrowding, and AI can contribute to that solution, says Ian Porée, former Executive Director of HM Prison and Probation Service, and Justice Advisor to NEC Software Solutions.

When the prison system came dangerously close to collapse in summer 2024, emergency early release measures eased the crisis, but they didn’t make the problem go away. 

The priority must now be to tackle the root causes of offending by supporting the rehabilitation of offenders in the community. But with justice teams’ workloads at an all-time high, practitioners are finding it harder than ever to deliver the high quality, targeted, face-to-face support that helps turn lives around.

These are very human problems, but could the intelligent use of AI help turn the tide for the justice system?

Government support for AI in justice

In July 2025, The Ministry of Justice published its very first AI Action Plan for Justice, a three year strategy to embed AI across all areas of the justice system. 

Lord Timpson, the ministry’s AI lead, expressed his enthusiasm for the strategy, having seen ‘real opportunities for AI to improve the working lives of our frontline staff and colleagues.’

The frontline staff Timpson mentions are fundamental to reducing the prison population and helping people break away from the cycle of crime. If AI can save time and provide insight from data, justice teams can use their very human skills to keep people out of prison.

All this is possible today, not sometime in the distant future. And here’s a few ways AI can help: 

As caseloads pile up, so do the banks of data that go with them. One of the most effective uses of AI is to help make sense of that data.

AI technology is capable of taking ten years of records, or 20,000 case files and providing a digestible, by humans, summary in next to no time. Something that would take a human many weeks to achieve on their own. By delegating admin heavy tasks to AI, probation officers get the facts they need, so they can focus on the people they support. 

Likewise, every conversation is important when you’re helping someone at risk of offending, but the human memory can’t recall each tiny, yet potentially crucial snippet from a meeting. 

AI can not only transcribe meetings, it can provide detailed summaries, using terminology, situations and vocabulary which are specific to the justice system, enabling teams to gain more insight from interactions with offenders. 

As long as practitioners are open about how they are using AI to speed up processes, people won’t worry they will be judged by a machine. Instead, they will benefit from more tailored support and we can protect the system where human professionals make the decisions.

People on probation are responsible for doing the hard work to change their behaviour. Just one poor choice, can set someone back. Something simple such as someone on probation missing an anger management treatment session or moving into a household with people of concern to the police.

AI can learn from situations which have led to people offending in the past, and use this intelligence to predict potential future risk which could otherwise remain hidden. Having this information early enables a probation officer to check in with the individual and help them make more positive choices.

Supporting an individual in this situation calls for very human skills. At a time of spiralling workloads, AI can help by using data to assist with triaging, so case workers can see which people out of their heavy caseload need the most urgent support or enforcement action. 

Identifying an individual who is at risk of criminal behaviour is critical. It signals the need to intervene before a person becomes a real risk in terms of public safety, or gets caught up in dangerous, damaging circumstances like county lines activity.

There are vast stores of data held by different organisations, such as policing, health, education, social housing and social work teams which AI can learn from. It could be used to better protect the public by to highlighting people at risk of harmful behaviour.

AI thrives in rules-based environments, so tools can be trained to sift through representative datasets, mitigate human bias and comply with information-sharing protocols, to improve data sharing between public services. 

In the justice sector, every decision has an impact on a person’s life, and sometimes that impact can be far-reaching. AI should never take on the role of decision-maker. Instead, the technology can provide objective, evidence-based information on which humans can base their decisions.

The prison population crisis may have been narrowly avoided this time, but we are not out of the woods yet. Overcrowded, unsafe conditions fail to achieve one of the prison service’s most important missions, to protect the public. 

The time has come to use the power of AI to help justice workers devote their very human skills to help people turn away from crime and rebuild their lives.

To find out more, download the new white paper from NEC Software Solutions: The role of AI in the human world of Justice