More than a decade of funding shortages by successive governments have put the UK's judicial system in jeopardy. Currently, if the court date has been cancelled due to logistical issues, there is an important backlog.
The powerful voices in British politics, including the Tony Blair Institute and Policy Exchange Think Tanks, put weight behind artificial intelligence (AI) as a potential solution to problems experienced across the public sector. Some of these voices believe that AI can free staff from bureaucratic workloads and give them more time to focus on the human aspects of justice, such as face-to-face engagement with clients.
In January, the Labour Government announced plans to “unleash” AI across the UK to “turbocharge” growth, boost living standards and revolutionize public services.
So how does AI affect the UK's judicial system?
The current focus on AI is driven primarily by the development of large-scale language models (LLMS). This is the technology behind AI chatbots such as ChatGpt. However, automation, machine learning, and other AI tools are not novel features of the judicial system.
Older tools such as Technology Assistance Reviews have used AI forms to allow lawyers to predict the relevance of documents to a particular case or issue. More debate is that in the case of probation and immigration, risk score algorithms are used.
Critics in the last example warn that these systems will entrench inequality and affect people who change their lives without knowledge.
However, these automated risk scoring systems are essentially substantially different from LLMS-based productivity tools aimed at streamlining management processes. The latter can draft statements as well as scheduling and transcription of meetings.
You can also obtain and summarise document reviews and case law sources. The obvious success stories include the old Bailey saving £50,000 by using AI to handle a summary of the litigation evidence.
How and why these tools are implemented – institutional context – is extremely important. When digital tools are used to reduce costs rather than provide more space for the human aspect of justice, the harm is especially prone to vulnerable clients.
This is because even these seemingly routine management uses of AI require human reviewers to catch plausible but incorrect information in order to cite the information generated by these tools and to exercise expert judgment.
Evidence from the small home office pilot scheme shows why this is important. The pilot scheme used LLMS to support asylum decisions by summarizing asylum case documents and transcripts.
Approximately 9% of the results were found to be inaccurate and lack of interview references. Another 23% of users testing the scheme were completely unsure about the summary despite significant time savings.
Justice and digitalization
In July 2025, the Ministry of Justice published its AI Action Plan for Justice. Microsoft's Copilot Chat is already available to law firm owners, but the strategic document has pledged to deploy AI tools to 95,000 justice staff by December.
The plan acknowledges many limitations on AI. It also set up the Chief AI Officer and creates AI guidelines, emphasizing that AI “does not support it, not support it” for human judgment.
It highlights careful approaches to rollouts, including efforts to gather feedback from trade unions and the public. It also emphasizes transparency through new websites and ethics frameworks.
The plan continues to promote more controversial use of technology, including assessing the risk of violence in detention. Nevertheless, it focuses more on LLMS to save on administrative tasks.
But could new strategies lead to the adoption of LLM tools by the judicial system before there is a mature understanding of how they are applied? Decisions based in part on AI-generated evidence are likely to provide new grounds for complaints and challenges. This could add, rather than reducing the case backlog.
In June 2025, a senior UK judge warned lawyers about the use of LLM tools as these tools could “hastised.” Elsewhere around the world, there have been many lawsuits that appear to have been filed in court for fictional AI-generated materials.
Given their limitations, the advantages of these tools are generally seen in the parts of the systems that are most resource- and time-consuming for human monitoring. Risk is the biggest hit when humans have low time and resources, and when clients have little money and time to challenge their decisions.
This unequal access to justice is not just about AI issues. The previous wave of digitalization used to reduce bureaucratic loads allowed several guilty pleas online and online convictions for several crimes automatically.
Gemma Burkett, a lecturer in criminal justice at Citi St. Georges University, argues that these automated systems affect particularly marginalized women and are far more likely to plead guilty to crimes they did not commit.
Fill over the cracks
There is a strong debate in favour of using bespoke, carefully developed technology to remove the administrative burden of justice system staff.
However, if your current system is struggling, adopting LLMS (or other forms of rapid digitization) does not solve the deep underlying problems caused by years of austerity. Rather than reducing the bureaucracy, they risk painting papers over cracks in dysfunctional systems.
