Karen Eltis
As synthetic content generated by artificial intelligence increasingly appears before the courts, Professor Karen Eltis, from the Civil Law Section, is joining a research team at the Canadian Institute for Advanced Research (CIFAR) that will develop a specialized tool to detect it.

This large-scale project received $700,000 in funding from the CIFAR Canadian AI Safety Institute Research Program at CIFAR (CAISI) to support its research and development efforts.

Synthetic content includes, among other things, fabricated images or videos, as well as court documents written using large language models like ChatGPT, which may contain inaccurate information and create numerous additional challenges.

“The growing sophistication of AI-generated content poses significant challenges for justice systems, both individually and institutionally, as the integrity of evidence and the independence of the judiciary are paramount,” notes Professor Eltis.

“Recent cases, including in Canada, highlight the urgent need to act in order to maintain public trust and support access to justice in the digital age,” she adds.

A Free Tool Developed in Collaboration with the Legal Community

CIFAR’s research team, known as a « Solution Network », will design and deploy a tool to verify synthetic evidence and documents, built specifically for the justice system and powered by AI. The tool will be offered for free, and its source code will be openly accessible.

The Solution Network will collaborate closely with lawyers and judges to ensure that the tool is aligned with procedural rules and the realities of the courtroom.

The project is co-led by Ebrahim Bagheri, Professor at the University of Toronto, and Maura R. Grossman, Professor at the University of Waterloo and at Osgoode Hall Law School. Alongside Professor Eltis, the team includes researchers from the University of British Columbia, Western University, and the University of Waterloo.

Another Solution Network has also been created by CIFAR to mitigate dialect biases and linguistic inequalities embedded in large language models.

Grounded in innovation and interdisciplinary collaboration, this project demonstrates how scientific research can generate concrete solutions to contemporary challenges. Professor Eltis and the CIFAR team show how research conducted within the Civil Law Section contributes to a more reliable and better-equipped justice system for the future.