The Law Society is urging stronger oversight of agentic artificial intelligence in the justice system, warning that the technology may develop faster than the rules needed to govern it. Its foresight study says systems that act on behalf of people could enter legal processes before society settles where automated action must stop and human judgment must begin.
The warning comes as court hearings are listed as far ahead as 2030 and the justice system faces what the Society calls “immense pressure.” More people are turning to artificial intelligence tools to navigate proceedings, but the Society said inaccurate or misleading information could create additional work for courts.
The Society distinguishes agentic, or autonomous, AI from generative AI. Generative systems create content, while agentic systems can take actions for users. That capability could create opportunities, but the Society said it also carries “significant risks” for justice, public services and government.
The study, The Future of Agentic AI in the Justice System, examines how to preserve trust as the technology emerges. It says agentic AI “will not fix a failing justice system” and could reinforce existing weaknesses, including underfunding, poor data and paper-based processes.
Risks to Fairness and Accountability
The report also describes a “justice gap” between legal-sector organizations with resources to use AI effectively and those without them. The Society warned that unclear or unequal access to advanced tools could widen inequalities, while errors could spread quickly and responsibility for decisions could become fragmented.
Another concern is that agentic AI could speed up the reclassification of some judicial functions as administrative processes. The report says that could make changes harder to challenge. It also says existing regulation is “poorly suited” to agentic AI.
The Society said careful use might free judges and practitioners to focus on work that “genuinely requires human judgement.” It called for clear accountability, greater transparency about AI decision-making, meaningful human oversight and public debate before widespread use.
Law Society vice president Brett Dixon said fairness, accountability and public trust must not be sacrificed for unmonitored efficiency. He said AI could ease some administrative pressure but would not replace investment in the court estate or additional court staff.
Human Judgment Remains Central
“Legal judgement cannot be delegated,” Dixon said, referring to courts and the wider legal sector. He said professional expertise to shape, check, question, improve and approve AI outputs is needed, and argued that the key question is what should remain human.
Dixon said there is still time to shape AI boundaries before the market does so. The Society’s warning adds to coverage of technology firms entering legal services and broader debate about innovation and public trust. The Society’s call for oversight also connects with reporting on debates over AI regulation.
The study’s stated next step is public discussion of the safeguards and boundaries for agentic systems before they become widespread in justice. Read the Society’s warning for the full source report and comments.
