Meet the 2026–27 A2AJ Student Innovation Fellows
We are pleased to introduce the 2026–27 cohort of A2AJ Student Innovation Fellows. Each project builds on A2AJ's open legal datasets to advance access to justice in Canada, and each will release its code and data openly.
Sarah Kamoun — Auditing and Mitigating Intersectional Bias in Canadian LJP Systems
Sarah is a Master of Applied Science student in Electrical and Computer Engineering at Toronto Metropolitan University and a member of the NSERC CREATE Responsible AI program. Her research focuses on fairness in AI systems, with a particular emphasis on Legal AI.
The Project
Sarah's fellowship project aims to build a Canadian fairness benchmark for legal judgment prediction (LJP): systems that forecast how courts will decide case outcomes. The work begins by structuring A2AJ case law into a dataset suitable for this task, then constructing versions of the dataset in which identity characteristics (like gender and ethnicity) are varied to assess whether AI models change their predictions based on the plaintiff or claimant's characteristics alone.
As AI tools are increasingly explored across the justice system, from legal aid clinics and advocacy to case triage and self-represented litigants, the fairness of these tools becomes an important concern. A model that systematically misjudges cases involving equity-deserving groups risks reinforcing the very barriers that access-to-justice work seeks to remove. By providing a dataset and evaluation methodology tailored to Canadian jurisprudence, this project constitutes an open-source foundation for evaluating demographic bias in Canadian legal AI systems.
Erin Peterson — Mapping Credibility: An Open-Source Microcorpus of Credibility Findings in Canadian Refugee Appeal Division Decisions
Erin is a Juris Doctor candidate at the Lincoln Alexander School of Law at Toronto Metropolitan University. Before law school, she completed a Bachelor of Applied Science in Computer Engineering with a specialization in artificial intelligence and spent three years working in AI infrastructure and software engineering. She also runs Integrity AI, an AI governance consulting practice. Her work focuses on responsible AI implementation and using AI for social good.
The Project
Erin's fellowship project builds an open dataset of credibility findings drawn from Refugee Appeal Division decisions in A2AJ's collection. When a refugee claim is refused, the reason is most often that the decision-maker did not believe the claimant. The project identifies those moments in each decision and records what kind of concern was raised, what evidence it rested on, and how the appeal was resolved. The dataset, the code, and the methodology will all be released publicly for anyone to use or extend.
Credibility is the hinge on which most refugee refusals turn, yet it is one of the least visible parts of the process. Lawyers, clinics, and community organizations can read decisions one at a time, but have no practical way to see patterns across them, whether similar concerns are treated consistently or whether certain claimants face particular kinds of doubt. By making credibility reasoning observable at scale, the project gives advocates evidence to draw on and researchers a foundation for studying fairness in refugee adjudication, including a focused look at claims involving sexual orientation and gender identity.
Sabrin Saide and Matt Aydin — Collective Bargaining Compass
Sabrin Saide is a third-year JD candidate at Osgoode Hall Law School, and Matt Aydin recently graduated from Osgoode and is beginning his articles. They share interests in labour and employment law, legal technology, and improving public access to legal information. Their work is focused on making complex workplace rights easier to understand and developing practical, user-centred tools for the broader public.
The Project
Collective Bargaining Compass will create an open-source, searchable, and standardized dataset of Ontario public-sector collective agreements. Collective agreements contain important rights related to wages, benefits, scheduling, leave, workload, job security, accommodation, and workplace protections, but they are often scattered across different websites and difficult to search or compare. Using legal review and AI-assisted coding, the project will organize key provisions into a publicly accessible dataset that allows users to compare agreements across employers, unions, bargaining units, and bargaining cycles. The long-term goal is to create a framework that can expand to other sectors and jurisdictions across Canada.
Hamed Taherkhani — Cases Like Mine / Des Causes Comme la Mienne
Hamed is a Ph.D. candidate in the EECS Department at York University. His research lies at the intersection of software engineering and artificial intelligence, with a focus on improving the reliability, reasoning, and effectiveness of large language models for software engineering. His work spans automated test generation, repository-level code understanding, execution-aware reasoning, reinforcement learning for code intelligence, and AI agents for software development. He has authored several first-author publications in leading software engineering venues, and has several years of experience developing software systems in industry.
The Project
Cases Like Mine / Des Causes Comme la Mienne is a free, open-source, bilingual web application designed to help self-represented litigants in Canada better understand how cases similar to their own have been decided. Users describe their legal dispute in plain language, and the system uses hybrid search and AI-based factual similarity analysis to identify relevant Canadian court and tribunal decisions, summarize comparable cases, show outcome statistics, link applicable statutory provisions, and highlight arguments that have historically succeeded or failed. It focuses on high-need areas such as income security and federal benefits, human rights and discrimination, and selected civil disputes in Ontario, British Columbia, and federal courts.
Built using A2AJ's Canadian case-law and legislation datasets, the project also creates and openly releases a new structured dataset capturing case outcomes, remedies, holdings, statutes, and argument results.
Chong Tan — Charter Intervention Accelerator
Chong is a Juris Doctor graduate of Osgoode Hall Law School. She also holds a Master of Law degree from China, and spent over five years as an intellectual property lawyer in Beijing before pursuing her JD in Canada. Her work focuses on the intersection of legal technology and access to justice in constitutional litigation.
The Project
Chong's fellowship project aims to build open infrastructure to reduce the resource burden on legal clinics and pro bono counsel throughout the Charter litigation process. It starts from a proof of concept that mapped 402 Supreme Court of Canada Charter cases from 2011 to 2025, which revealed for the first time the network of clinics, NGOs, and counsel that shapes Charter intervention in Canada. The fellowship will develop this foundation into a broader toolkit that addresses the friction points at each stage of constitutional advocacy, from identifying cases to finding counsel to drafting intervention applications. The project will also extend this infrastructure beyond the Supreme Court to provincial courts of appeal.
Charter intervention shapes Canadian constitutional law, but the process favours organizations that already have resources and connections. The communities most affected by Charter decisions are often the ones least able to participate. This project tries to close that gap by making the pro bono capacity that already exists in the legal market more findable and more usable for the people who need it most.
Xiang Zhang — Legal Language Bridge
Xiang is a doctoral student at Osgoode Hall Law School at York University. He completed a Bachelor of Engineering in Software Engineering before law school. His research focuses on intellectual property, particularly on topics related to open-source and open access to information and knowledge.
The Project
Xiang's fellowship project, Legal Language Bridge, develops an open-source, multilingual search tool for A2AJ's legal datasets. Users will be able to describe a legal problem in plain language, and the tool will identify likely legal concepts, translate the description into Canadian legal terms for query, and return source-linked cases, legislation, and other relevant materials. The project is designed for self-represented litigants, community advocates, new immigrants, legal clinic users, and others who need legal information without knowing the specialized vocabulary required by conventional databases.
Legal information can be publicly available but still remain out of reach when people cannot translate their lived experiences into the terminology used by courts and legal databases. Language, cultural, and institutional barriers deepen this problem, especially for people who do not search comfortably in English or French. By turning plain language descriptions into legal terms and directing users to authentic legal materials rather than unverified legal advice, Legal Language Bridge lowers the vocabulary barrier to the justice system and helps A2AJ's legal data reach the people who may benefit from it most.
About the Program
The A2AJ Student Innovation Grants provide $10,000 per project, supported by the Law Foundation of Ontario. Fellows work with A2AJ's Canadian legal data to build open-source tools that tackle access to justice challenges.
We'll be sharing updates on each fellow's progress throughout the year. To learn more about the grants program, visit our Projects page.