The Justice Catalyst | June 2026

The Justice Catalyst provides intelligence on how AI is shaping justice and justice is shaping AI. In this issue: AI legal assistants for the UK Crown Court; India’s Supreme Court draft rules on AI in adjudication, sentencing, and risk assessment that explicitly invite assistive use; and Brazil’s National Council of Justice second-edition study on generative AI in the judiciary. The Strategic Reads section features seven 2026 empirical studies on AI in adjudication.

Download the June 2026 issue (PDF)

News

Policy and deployment

UK Ministry of Justice expands AI initiatives with an AI Growth Lab and Crown Court AI legal assistants

June 9, 2026

Justice Secretary David Lammy announced two new components to the Ministry’s AI initiative: the AI Growth Lab regulatory sandbox with an initial focus on legal services, and integration of AI legal assistants to support routine casework, research, and case analysis in the Crown Courts.

Sources: Hilborne, Legal Services Lead Way as Government’s First AI Growth Lab, Legal Futures (June 9, 2026), https://www.legalfutures.co.uk/latest-news/legal-services-lead-way-as-governments-first-ai-growth-lab; AI Legal Assistants to Help Speed Up Work of Crown Court, Legal Futures (June 9, 2026), https://www.legalfutures.co.uk/latest-news/ai-legal-assistants-to-help-speed-up-work-of-crown-court; Lord Chancellor David Lammy to Announce AI Legal Assistants, Law Gazette (June 9, 2026), https://www.lawgazette.co.uk/news/ai-assistants-could-help-cut-backlog-says-lammy/5127002.article; Regulators SRA and LSB Join Lammy’s AI Sandbox, Law Gazette (June 9, 2026), https://www.lawgazette.co.uk/news/we-want-to-be-early-adopters-justice-to-lead-governments-ai-action-plan/5127008.article; Ministry of Justice & Dep’t for Sci., Innovation & Tech., Legal Innovation to Be Supercharged by New AI Growth Project, GOV.UK (June 8, 2026), https://www.gov.uk/government/news/legal-innovation-to-be-supercharged-by-new-ai-growth-project.

India Supreme Court releases draft Regulations for Use of AI in Courts

June 4 to 5, 2026

The Supreme Court’s AI Committee released draft rules barring use of AI in judicial decision-making and bail-risk assessment as well as behavioral profiling and outcome prediction. At the same time, the rules explicitly encourage assistive AI use under an “Innovation Over Restraint” philosophy anchored in human primacy, data protection, and transparency. It invited public comment on the draft through June 20.

Sources: Regulations for Use of Artificial Intelligence in Courts, 2026, Verdictum (June 4, 2026), https://www.verdictum.in/supreme-court/regulations-for-use-of-artificial-intelligence-in-courts-2026-1615304; Draft AI Rules Mandate Disclosure of AI Use in Court Filings, The Federal (June 5, 2026), https://thefederal.com/category/news/supreme-court-draft-ai-rules-disclosure-ai-use-court-filings-245443; Supreme Court of India, Draft Regulations for Use of Artificial Intelligence in Courts, 2026 (June 4, 2026), https://www.verdictum.in/pdf_upload/2026/06/04/2026060342-2watermark-1778772.pdf; Mehta & Nayak, Regulating Artificial Intelligence in Indian Judiciary, LiveLaw (June 11, 2026), https://www.livelaw.in/articles/regulating-artificial-intelligence-indian-judiciary-537485.

Singapore Chief Justice Menon on AI

May 22, 2026

Chief Justice Sundaresh Menon announced a “traffic light” model for AI adoption in Singapore’s Courts.

Source: Truth Decay and AI in Courts, The Online Citizen (May 26, 2026), https://theonlinecitizen.com/2026/05/26/truth-decay-and-ai-in-courts-chief-justice-menon-on-challenges-facing-singapore-judiciary.

Brazil CNJ launches Sinapses 2.0 and second-edition generative AI study

April 24, 2026

The National Council of Justice launched Sinapses 2.0, a national artificial intelligence ecosystem designed to develop, train, store, version, and audit AI models in the Brazilian justice system, alongside the second edition of its study on generative AI use across the judiciary.

Source: CNJ, Lançamento de Inovações de Inteligência Artificial para o Judiciário (Apr. 24, 2026), https://www.cnj.jus.br/lancamento-de-inovacoes-de-inteligencia-artificial-para-o-judiciario-marcam-evento-no-cnj/.

Women Transforming Justice

Partnerships

Maya Markovich and Kalina Leopold Oak, Justice Technology Association

The Justice Technology Association, led by Maya Markovich (Executive Director) and Kalina Leopold Oak (Chief Operating Officer), partnered with Anthropic as the access-to-justice launch partner for Anthropic’s first comprehensive legal vertical initiative.

Source: Ambrogi, JTA Named Access to Justice Partner in Anthropic’s Legal AI Launch, LawSites (May 4, 2026), https://www.lawnext.com/2026/05/justice-technology-association-named-access-to-justice-partner-in-anthropics-legal-ai-launch.html.

Funding and milestones

Ebru Metin and Prof. Helena Haapio, European Legal Design Action

Ebru Metin (TalTech Law School) and Prof. Helena Haapio (University of Vaasa) saw their European Legal Design Action (ELDA) proposal selected for COST funding as one of 80 Actions from the highly competitive Open Call. The four-year program launches in autumn 2026 to advance legal design across Europe through annual conferences, training schools, working groups, and short-term scientific missions.

Source: European Cooperation in Sci. & Tech., 80 New COST Actions Approved Under Open Call OC-2025-1 (May 19, 2026), https://www.cost.eu/80-new-cost-actions-oc20251/.

Dorna Moini, Founder and CEO, Gavel

Dorna Moini, founder and CEO of Gavel, saw Gavel acquired by Relativity. Moini founded Gavel to support her work assisting domestic violence survivors and expanded it to support the legal aid community and practitioners more generally. The AI-native document automation platform, used by thousands of legal professionals, joined Relativity.

Sources: Relativity, Relativity Acquires Gavel, PR Newswire (June 12, 2026), https://www.prnewswire.com/news-releases/relativity-acquires-gavel-to-extend-its-ai-platform-for-legal-data-intelligence-into-microsoft-word-302799189.html; Relativity’s Legal AI Play Takes Aim at Microsoft Word, LawFuel (June 12, 2026), https://www.lawfuel.com/relativitys-legal-ai-play-takes-aim-at-microsoft-word/.

Recognitions

Amanda Bell, Founder and CEO, Separate Space

Amanda Bell, founder and CEO of Separate Space, was named Business Leader of the Year at the Women and Diversity in Law Awards 2026. Separate Space’s AI-powered platform delivers low-cost, plain-English, child-centric family justice and has helped more than 17,000 people since launching in 2024.

Source: 2026 Winners, Women & Diversity in Law Awards (May 9, 2026), https://womenanddiversityinlawawards.com/womenanddiversityinlawawards/en/page/winners-2026.

Strategic Reads

AI and access to justice

Access to Justice in the Age of AI: Evidence from U.S. Federal Courts

Anand Shah and Joshua Levy (USC), SSRN, March 20, 2026

Empirical study on the impact of AI on access to justice. Findings include that AI-generated text in pro se federal complaints rose from near zero to over 18% in 2026 and that pro se docket activity is up 158%. However, case resolution has not accelerated.

Source: Shah & Levy, Access to Justice in the Age of AI, SSRN abstract 6766859 (Mar. 20, 2026), https://ssrn.com/abstract=6766859.

Innovating for Access: AI-Enhanced Triage and Intake for Legal Services Organizations

Heidi Behnke, Duke Center on Law and Technology, March 16, 2026

This study examines AI use among legal services organizations including projects, infrastructural needs, and best practices for responsible adoption.

Source: Behnke, Innovating for Access, Duke Ctr. on L. & Tech. (Mar. 16, 2026), https://scholarship.law.duke.edu/dclt/3.

Access to What? Lawyers, Transaction Costs, and Justice in the Age of AI

David McGowan, San Diego Legal Studies, SSRN, February 20, 2026

This paper examines the question of what access to justice means when transaction costs collapse, concluding that reduced cost alone does not necessarily provide access and recommending new AI-powered solutions avoid structures that equate justice with the presence of a lawyer.

Source: McGowan, Access to What?, San Diego Legal Studies Paper No. 26-010, SSRN (Feb. 20, 2026), https://ssrn.com/abstract=6276898.

Governance and frameworks

An AI Taxonomy for Criminal Justice

RAND and Council on Criminal Justice, May 15, 2026

This paper provides a rubric for AI applications by function and decision point across law enforcement, courts, corrections, and supervision.

Sources: Council on Crim. Just., An AI Taxonomy for Criminal Justice, RAND Corp. (May 15, 2026), https://counciloncj.org/wp-content/uploads/2026/05/AI-Taxonomy.pdf; RAND Corp. Publication Record (2026), https://www.rand.org/pubs/external_publications/EP71334.html.

Judicial AI and the Bottleneck Shift

bo liu, SSRN, April 24, 2026

This report conducts a role-level analysis across 56 court occupations and concludes that AI is most likely to drive shifts in roles involving judicial support and other administrative functions.

Source: bo liu, Judicial AI and the Bottleneck Shift, SSRN (Apr. 24, 2026), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6612398.

AI deployment

AI on Trial: 388 decisions track AI references across Canada’s courts

Canadian Lawyer data report, March 2026

This report examines treatment of AI in 388 Canadian legal decisions (2021 to mid-2025) across 77 courts and tribunals and includes expert commentary on Canada’s approach.

Source: AI on Trial: Federal Court, Canadian Lawyer (Mar. 2026), https://www.canadianlawyermag.com/cl-plus/data-reports/ai-on-trial-388-decisions-track-the-rise-of-ai-references-across-canadas-courts-federal-court/393875.

AI and digital justice in EU labor law: a comparative study on predictive tools and judicial transformation

Marianna Molinari (Bologna) and Marco Giacalone (Turin and VUB Brussels), Frontiers in AI, January 29, 2026

This comparative study reviews use of AI-enabled predictive tools and judicial transformation across Belgium, Croatia, Czech Republic, Estonia, Italy, and Lithuania.

Source: Molinari & Giacalone, AI and Digital Justice in EU Labor Law, Frontiers in AI (Jan. 29, 2026), https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12894374/.

Artificial Intelligence in Brazilian Courts, 2019 to 2026: a scientific analysis

Revista de Direito e Tecnologia, 2026

This report traces AI implementation in the Brazilian Judiciary focusing on efficiency outcomes and ethical and regulatory implications.

Source: Inteligência Artificial nos Tribunais Brasileiros, Revista de Direito e Tecnologia (2026), https://remunom.ojsbr.com/multidisciplinar/article/view/5940.

AI in adjudication

Seven 2026 studies evaluate AI in adjudication, contributing to a growing body of peer-reviewed research on the topic.

Persuadability and LLMs as Legal Decision Tools

Oisin Suttle and David Lillis, arXiv, June 8, 2026

Study of how persuadable AI models are when given hard legal questions from real cases. The researchers designed an experiment using real appellate court cases from Ireland, England and Wales, and the United States, pitting AI models against each other with one as Advocate, another as opposing Advocate, and a third as Judge. Across 20 model configurations from Anthropic, Google, OpenAI, Mistral, and DeepSeek, stronger advocate models won 58 to 71 percent of cases on average, exceeding 90 percent in some cases. The study found that sophisticated legal representation and “more powerful” AI advocates have a systematic advantage. Presented at ICAIL 2026 in Singapore.

Sources: Suttle & Lillis, Persuadability and LLMs as Legal Decision Tools, arXiv:2604.26233 (June 8, 2026), https://arxiv.org/abs/2604.26233; Carolan, Research Raises Concerns About Impact on Justice Systems of Using AI, Irish Times (June 8, 2026), https://www.irishtimes.com/crime-law/2026/06/08/research-raises-concerns-about-impact-on-justice-systems-of-using-ai/.

Evaluation of Large Language Models in Legal Applications

arXiv, January 2026

This study examines model performance on different law-related tasks including judicial applications.

Source: Evaluation of LLMs in Legal Applications, arXiv:2601.15267 (Jan. 2026), https://arxiv.org/abs/2601.15267.

Evaluating Large Language Models as Judicial Decision-Makers

Taylor & Francis, January 21, 2026

This study evaluates performance of GPT, Gemini, and Claude vs. 123 retired judges in sentencing. Models produced significantly lower sentence disparity than the judges.

Source: Evaluating LLMs as Judicial Decision-Makers, Taylor & Francis (Jan. 21, 2026), https://www.tandfonline.com/doi/full/10.1080/07418825.2026.2618254.

Large Language Models in Judicial Assistance: Empirical Insights and Domain-Specific Fine-Tuning

Natural Language Processing, Cambridge Univ. Press, February 2026

This study tests 13 mainstream models across five judicial tasks including citation, decision generation, and opinion generation.

Source: LLMs in Judicial Assistance, NLP, Cambridge Univ. Press (Feb. 2026), https://www.cambridge.org/core/journals/natural-language-processing/article/large-language-models-in-judicial-assistance-empirical-insights-and-domainspecific-finetuning/C3F61269E92774BACA37B01C840AAD0E.

Judge AI: A Case-Study of Large Language Models as Judges

Eric A. Posner and Shivam Saran, SAGE, March 19, 2026

This study examines the judicial ability of OpenAI’s GPT-4o and finds it is a competent judge who applies precedent correctly. However, its performance lacks sympathy.

Source: Posner & Saran, Judge AI, SAGE Journals (Mar. 19, 2026), https://journals.sagepub.com/doi/10.1177/2755323X261433614.

Assessing Cognitive Biases in LLMs for Judicial Decision Support

arXiv, March 2026

This study examines bias in model performance in judicial contexts.

Source: Assessing Cognitive Biases in LLMs for Judicial Decision Support, arXiv:2603.10016 (Mar. 2026), https://arxiv.org/abs/2603.10016.

TriBench-Ko: Evaluating LLM Risks in Judicial Workflows

arXiv, May 2026

TriBench-Ko is a Korean benchmark designed to evaluate potential deployment risks of LLMs within the context of judicial tasks. It covers four core tasks: jurisprudence summarization, precedent retrieval, legal issue extraction, and evidence analysis. It jointly assesses model behavior across multiple deployment risk categories.

Source: TriBench-Ko, arXiv:2605.03792 (May 2026), https://arxiv.org/abs/2605.03792.

Opportunities and Events

Learning and networks

AI, Justice and Rule of Law: free global MOOC (UNESCO and University of Oxford)

English course open since April 27, 2026; French and Spanish editions June 2026

A free global course on AI and the rule of law, jointly launched by UNESCO and the University of Oxford.

Source: UNESCO, UNESCO and University of Oxford Launch Free Global Course on AI and the Rule of Law (Apr. 27, 2026), https://www.unesco.org/en/articles/unesco-and-university-oxford-launch-free-global-course-ai-and-rule-law.

European Legal Design Network: applications open

Applications to join the Network are now being accepted.

Apply: European Cooperation in Sci. & Tech., https://www.cost.eu/.

Get involved

Join a Justice Circle

A peer advisory group coaching program designed to accelerate the leadership and success of women leading justice tech, tech nonprofits, and legal aid innovation, through structured peer learning and executive coaching. Circles support women entrepreneurs building market-based tech and nonprofit leaders building tech solutions to expand access to law and justice.

Each Circle is a ten-month cohort of 6 to 8 executive peers guided by a Lady Justice Initiative Lex 2X Exec™ credentialed coach.

Apply to join →

Call for Submissions

Send opportunities, milestones, and events for inclusion in future issues to LJI@ladyjusticeinitiative.org.

Women Building the Future of Justice

Submit a profile to be included in our dashboard, co-produced with Justice Technology Association. Share your company and individual achievements.

Submit your profile →

Stay Connected

Follow Lady Justice Initiative on LinkedIn for updates and insights at the intersection of AI, law, and justice.

Support Our Work

Donate and help build an AI era of law where women thrive and justice is native. Donate →