The Modernising Criminal Justice 2026 conference opened with Fiona Rutherford, Chief Executive of the UK legal reform charity JUSTICE, setting out why technology - and specifically AI - needs to be handled with care across the justice system.
Fiona explained that JUSTICE has spent nearly 70 years working to build a fair UK justice system, bringing together voices from government, policing, academia, tech and law. She used her opening address to share three questions her organisation applies whenever it assesses AI in a justice context:
1. Does it advance a genuine goal? Does the tool improve access to justice, decision-making or public scrutiny - or is it being used simply because the technology exists?
2. Where are the risks? These range from biased data and damaged public trust to poor value for money, and they exist at technical, individual and system-wide levels.
3. What's the responsible response? What steps will reduce those risks to a proportionate level, and can problems be prevented rather than fixed after the fact?
Fiona then shared findings from JUSTICE's recent work on AI in policing. Three themes came up repeatedly: the importance of choosing a tool built for the specific context rather than adapting one designed for something else; the need for informed, empowered human oversight; and the value of public participation in designing systems that communities can trust.
The headline gap her team identified is structural. England and Wales have 43 police forces, each currently free to experiment with AI on its own terms - setting its own standards and making its own decisions on transparency. Fiona's recommendation is a central body to set national rules for AI use in policing, working alongside a citizens' panel so the public has a genuine say in how future policing looks. She noted that the Home Office and other government departments have engaged constructively with this proposal, and she is hopeful it will feed into policy reform.
Fiona was clear that the challenge doesn't stop at policing. With criminal courts under severe strain and facing calls for radical reform, she said AI has a role to play there too - and flagged JUSTICE's ongoing research into how processes like disclosure could use AI fairly, provided the right safeguards are in place.
She closed on a sober note, pointing to the previous week's riots in Southampton as a reminder of what happens when public trust in the justice system breaks down. Her message to the room was that everyone present has a responsibility to use technology - AI included - in ways that reinforce rights and the rule of law, not erode them.
Fiona's framing is a useful checklist for anyone currently evaluating or procuring AI tools: ask whether it serves a genuine purpose, map the risks at every level, and build in proportionate safeguards from the outset. Her call for a central, cross-force standard-setting body is also worth watching - if it gains traction, it could reshape how individual forces and agencies are expected to govern their own AI use in future.