The push to adopt Artificial Intelligence (AI) tools in the UK justice system, while aimed at efficiency, runs the significant risk of masking deep-seated issues caused by more than a decade of government underfunding, which has led to massive case backlogs and logistical crises. The rapid adoption of AI to handle bureaucratic tasks may provide short-term cost savings but ultimately fails to address the underlying structural resource shortages and could severely undermine fairness and public trust.
The Crisis Driving AI Adoption
The primary driver for the UK justice system’s interest in AI, particularly Large Language Models (LLMs) like those behind chatbots, is the desperate need to address the operational crisis caused by long-term austerity measures. Years of underfunding have led to a significant case backlog, cancelled court dates, and reduced staffing. Proponents of AI argue that it can “turbocharge” public services by automating routine, bureaucratic tasks—such as document review, summarizing case law, and managing administrative workloads—thereby freeing up human staff to focus on the essential human aspects of justice, like face-to-face engagement with clients. Apparent successes, such as the Old Bailey reportedly saving £50,000 by using AI for evidence overviews, fuel this narrative.
The Risk of Entrenching Inequality
While administrative uses of AI appear benign, the context of their implementation is crucial. When digital tools are implemented primarily to cut costs rather than enhance human capability, the harms disproportionately fall on the most vulnerable clients.
Older AI forms, such as risk-scoring algorithms used in probation and immigration cases, have already faced criticism for entrenching existing societal inequalities and affecting people’s lives without their knowledge. When new LLMs are introduced into a system with severely limited human resources for oversight, the risk of error, bias, and injustice is magnified. Marginalized groups, who often lack the money and time to challenge decisions, are hit the hardest when human expert judgment is minimized in favor of automated cost-saving measures. This unequal access to justice is a pre-existing flaw of the system that AI tools threaten to exacerbate.
The “Hallucination” Danger and Erosion of Trust
A significant and immediate risk of adopting generative AI in the legal field is the potential for the technology to “hallucinate,” or generate completely fictitious information. Senior UK judges have already issued warnings to lawyers after international and local cases emerged where AI-generated fake case-law citations were filed in court.
Relying on AI without strong human oversight risks a new wave of legal challenges based on erroneous or misleading evidence, which could ironically add to, rather than reduce, the case backlog. More fundamentally, the misuse of AI—even if inadvertent—has serious implications for public confidence in the justice system. The system’s legitimacy rests on its ability to uphold the rule of law fairly. When digital tools fail or introduce demonstrable errors, it risks fraying social cohesion and deepening public mistrust between citizens and the state.
The Imperative of Human-Centric Reform
The current crisis facing the UK justice system is one of resource allocation, not technological deficiency. The government’s new AI action plan for justice rightly emphasizes that AI should “support, not substitute” human judgment and stresses a need for transparency and ethics. However, to truly benefit from AI, the system needs to prioritize a rights-based framework that ensures the technology is only used to improve the core goals of justice: access to justice, fair decision-making, and transparency. Without first addressing the deep, underlying problems of underfunding and resource depletion, adopting LLMs will not revolutionize the justice system; it will simply serve as a temporary patch over a fundamentally dysfunctional structure, ultimately delaying the necessary structural reform.
