TLDR: The Access to Algorithmic Justice (A2AJ) project introduces an open-source Canadian Legal Data initiative, providing free and programmatic access to over 116,000 court decisions and 5,000 statutes. This project aims to counter the Canadian Legal Information Institute (CanLII)’s restrictive data access policies, which are creating a digital divide in computational law, disproportionately affecting low-income communities and hindering legal innovation. A2AJ offers various access methods, including APIs and machine learning datasets, to foster a more equitable and accessible justice system.
In an increasingly digital world, access to information is power. This holds especially true in the legal sector, where new computational techniques like artificial intelligence are poised to transform how we understand and interact with the law. However, a recent working paper highlights a growing concern in Canada: a digital divide in legal information access, largely due to the restrictive policies of the Canadian Legal Information Institute (CanLII).
The paper, titled “Access to Algorithmic Justice Working Paper Introducing the A2AJ’s Canadian Legal Data: An open-source alternative to CanLII for the era of computational law,” authored by Simon Wallace and Sean Rehaag, introduces a groundbreaking initiative: the Access to Algorithmic Justice (A2AJ) project. This project aims to provide an open-source alternative to CanLII, making Canadian legal data freely and programmatically accessible to all. You can read the full paper here: Access to Algorithmic Justice Working Paper.
The Historical Context of Legal Data Access in Canada
The struggle for open access to legal information in Canada is not new. Historically, access to vast collections of caselaw was often limited to those who could afford proprietary databases. While the establishment of CanLII in 2001 by the Federation of Law Societies was a significant step towards democratizing access, its current policies are creating new barriers.
The paper traces a pattern where academic innovations in legal technology, such as Queen’s University’s QUIC/LAW project in the 1960s, were often commercialized, leading to restricted access. Even initiatives like the Canadian Law Information Council (CLIC) in the 1970s struggled to create a public search tool due to resistance from private publishers unwilling to share their databases.
CanLII: From Open Access Champion to Gatekeeper?
CanLII was initially celebrated for disrupting the market by offering free online access to millions of Canadian legal documents. However, the paper argues that CanLII is now impeding legal innovation by prohibiting bulk and programmatic access to its extensive dataset. This means that while well-resourced entities can afford expensive legal tech built on proprietary data, public and low-income communities are left with second-rate tools, if any.
The authors point to CanLII’s restrictive terms of use and copyright claims as key obstacles. They also note the evolution of LexUM, the technology supplier that built CanLII, from an academic project to a private company, eventually acquired by CanLII. This shift, the paper suggests, has led to a focus on revenue generation and a reluctance to share data widely, even as LexUM uses CanLII’s data to develop new AI technologies for commercial partners.
A2AJ’s Canadian Legal Data: An Open-Source Solution
In response to these challenges, the A2AJ project has developed its own Canadian Legal Data (CLD) initiative. This open-source alternative currently provides access to over 116,000 court and tribunal decisions and more than 5,000 statutes and regulations, many available in both English and French. This amounts to over 1.2 billion tokens of legal text, making it a substantial resource for researchers and developers.
The A2AJ obtains its data through automated scraping of public court and tribunal websites, direct agreements with administrative tribunals, and from the Department of Justice for federal legislation. Crucially, the project is committed to complying with the terms of service of its data sources, and it passes these upstream licenses along to its users.
Multiple Pathways to Access
Recognizing diverse user needs, the A2AJ offers several free and registration-free methods to access its data:
API Access: For targeted searches and integration into applications, allowing programmatic queries by citation, name, full text, date ranges, and jurisdiction.
Hugging Face Datasets: For machine learning researchers, the bulk datasets are hosted on Hugging Face, simplifying the process of loading data for training language models and computational legal studies.
Direct Parquet Downloads: For large-scale empirical analysis, the entire dataset can be downloaded in the efficient Parquet format, compatible with most data analysis tools.
Model Context Protocol (MCP): For AI integration, enabling AI assistants and chatbots to query the legal database in real-time, providing accurate and up-to-date Canadian legal information.
Real-World Applications of Open Legal Data
The paper provides compelling examples of how A2AJ’s open legal data can be used:
Supporting Efficiency and Accessibility in Courts: Courts can use the CLD to conduct data-driven assessments of their own practices. For instance, the Tax Court of Canada could measure the impact of readability training by analyzing decision texts, or the Federal Court could identify judges who consistently produce concise decisions in high-volume areas like immigration judicial reviews.
Automated Decision Summarizers and Podcast Generators: For overworked legal professionals, the CLD enables the creation of tools that can automatically summarize weekly jurisprudence, identify errors in decisions, and even transform these summaries into podcast scripts. The paper highlights how a basic version of such a tool was built in just two hours, demonstrating the ease of development with A2AJ data.
Also Read:
- JustEva: A New Toolkit for Assessing Fairness in Legal AI Models
- A Structured Approach to AI Governance: Bridging Regulation and Practical Implementation
A Call for Openness
The A2AJ project underscores the urgency of democratizing access to legal data in Canada. The authors argue that while CanLII was a major achievement, it needs to reclaim its leadership in the era of computational law by making its data freely available for computational use. They call upon law societies, the Federation of Law Societies, and courts and tribunals to recognize legal data as a public good, not a privatized asset, and to support initiatives that ensure equitable access to justice for all Canadians.
