Yvon Dandurand

Pathways to justice research has examined the full range of everyday legal problems, the paths individuals take to address them, the obstacles they face, the assistance they seek or receive, and the outcomes of their choices and actions. [i]–[ii] Knowing the pathways people take is essential in making sense of whether and how their legal needs are met.[iii]  Pathways to justice are differentially accessible to individuals facing legal problems, depending on the nature of the problem, the private costs of pursuing a particular path (including uncertainty about those costs), and other obstacles that, depending on the situation and other factors, affect individuals differently.[iv] When people face a justiciable problem, they may or may not pursue a path, or problem-resolution route, to justice.[v] Agentic AI opens not just a new path, but a new highway. AI has already begun to reshape people’s pathways to conflict resolution and access to justice.

The transformation is unfolding so quickly that empirical research cannot keep pace, and access-to-justice policies become obsolete before they can be fully implemented. One of the few available studies examined evidence of agentic AI’s impact on U.S. Federal Courts.[vi]

It documented how generative AI has reshaped entry into the federal civil court system. It also found that the number of pro se (self-represented) cases is increasing dramatically, rising from a long-term steady-state average of 11% to 16.8% in FY2025 and placing a greater burden on federal district courts.[vii] It also showed that a large and growing share of complaints flag positive for AI-generated text, rising from essentially zero in the pre-AI period to more than 18% in 2026. It also found that this rise concentrates in cases where document production is the primary cost of filing, and that it has occurred almost uniformly across states. In brief, these technological advances have materially altered filing behaviour.

Earlier this month, the Law Society (UK) published The Future of Agentic AI in the Justice System, examining how increasingly autonomous AI systems could reshape courts, tribunals and dispute resolution in England and Wales. The report notes that legal AI has rapidly evolved from a tool that answers questions into an agent that takes action: “Over the next decade, increasingly capable agentic AI systems may become the first point of contact between citizens and the justice system—helping people identify legal problems, decide whether to act, prepare claims, negotiate outcomes and manage interactions with courts and legal services.”[viii] In other words, the next decade may determine whether agentic AI becomes the greatest expansion of legal capability in history, or creates a new layer of invisible gatekeeping between people and justice.

The report concludes that a series of seemingly incremental decisions about technology and system design could gradually alter not only how justice is delivered but also where authority sits within the system. It warns that “the new gatekeepers of justice may not sit in courts or law firms, but inside technology platforms. If commercial AI systems become the main route through which people understand their rights and navigate disputes, access to justice could become shaped by private incentives, business models and invisible design choices.”[ix] Indeed, agentic systems provided by technology platforms, insurers, employers, and large institutions increasingly prevent, manage, negotiate, and resolve disputes. These systems offer speed, convenience and lower cost.[x] For many people, courts are already no longer the primary gateway to justice.

Meanwhile, ironically, a rising wave of AI-generated claims threatens to swamp governments, tribunals, and courts. More people now have the power to access justice and conflict-resolution institutions. They use AI to access a wide range of skilled capacities previously held by only a small number of specialists. More people can learn about their legal rights from AI and use it to file more claims for legal remedies. However, where generative AI is used excessively or inappropriately, it can create challenges. We all know examples of submissions using generative AI that are unduly lengthy, incoherent, or contain ‘hallucinations’ (misleading, inaccurate, or non-existent information or references to laws, regulations, and previous adjudications).

Evidence also suggests that claims management companies and legal representatives use AI to generate submissions on behalf of their clients. When they are charging significant fees for such content, this is especially concerning. As the Law Society of BC noted, adopting generative AI tools offers lawyers exciting opportunities for efficiency and productivity but also introduces ethical challenges that demand vigilance.[xi]

Schmitz, Hammond and Chan[xii] have identified several potential government responses to Agentic Flooding. For example, governments could suppress demand for services, as they have done in the past, by making applications harder or less attractive to complete, or they could raise submission costs until enough users are deterred from filing, thereby suppressing quantitative flooding. Governments could also lower what a user stands to gain from submitting to a service. Alternatively, governments could respond to flooding by increasing their processing capacity or redesigning services. Obviously, governments could use AI agents themselves– both to accelerate back-end processing and to detect and prevent fraud. Governments are already taking first steps in this direction. No precedent exists for more comprehensive agent deployment, but evidence suggests it could drastically scale processing throughput. Some countries already use AI tools for intake screening, triage, drafting standard correspondence, and even autonomously handling routine cases.

The speed at which agentic AI is empowering people and transforming access to justice and conflict-resolution institutions points to the need for a concerted effort by governments and civil society to respond constructively to this transformation. I suggest that an independent study is urgently needed to monitor agentic AI’s drastic impact on our courts and regulatory institutions.

[i] Capp, M. & Dandurand, Y. (2024). Towards People-Centered Justice: The Conflict Resolution Routes of People Facing Legal Problems. Canadian Journal of Law and Society, 39:3, 468–489.

[ii] Dandurand, Y., Jahn, J., Tait, C., & Capp, M. (2022). Navigating access to justice pathways: Problem resolution routes for people experiencing civil and family law problems in British Columbia. University of Victoria Access to Justice Centre for Excellence.

[iii] Access to Justice BC (2017). Access to Justice Measurement Framework. A project funded by the Law Foundation of BC. https://drive.google.com/file/d/15gtf7TpqcTofY3XIyGR-BeK74CXIteLq/view

[iv] Gramatikov, M., Barendrecht, M., & Verdonschot, J. H. (2011). Measuring the costs and quality of paths to justice: Contours of a methodology, Hague Journal on the Rule of Law, 3(2), pp. 349-379.

[v] Gramatikov, M. (2009). A Framework for Measuring the Costs of Paths to Justice. The Journal Jurisprudence, 2, p. 111, TISCO Working Paper Series on Civil Law and Conflict Resolution Systems No. 006/2008

[vi] Shaw, A. V. & Levy, J. Y. (2026). Access to Justice in the Age of AI: Evidence from U.S. Federal Courts. https://ssrn.com/abstract=6766859 or http://dx.doi.org/10.2139/ssrn.6766859

[vii] Shaw & Levy, Access to Justice in the Age of AI, P. 10

[viii] The Law Society (2026). The Law Society Foresight Report: The future of agentic AI in the justice system: Implications for courts, tribunals and alternative dispute resolution, p. 29, https://www.lawsociety.org.uk/topics/research/the-future-of-agentic-ai-in-legal-practice

[ix] The Law Society Foresight report, p. 30.

[x] The Law Society Foresight report, p. 77.

[xi] Law Society of British Columbia (2023). Guidance on Professional Responsibility and Generative AI. https://www.lawsociety.bc.ca/getContentAsset/91264ae3-9fa0-4063-88fb-69977a293f62/dfc3d011-8f63-43f6-9ed8-4b444333a1d0/Professional-responsibility-and-AI.pdf?language=en-CA

[xii] Schmitz, C., Hammond, L., & Chan, A. (2026). Characterizing Agentic Flooding of Government Services. Proceedings of the 9th AAAI Conference on AI, Ethics, and Society (AIES), October 12–14, 2026

Photo by Igor Omilaev on Unsplash.