Can generative AI be useful in legal research?
Artificial intelligence has long been present in legal research systems—from natural language search engines to enhanced citators for noting up. Generative AI, however, is a fairly recent development.
Generative AI refers to AI systems that produce new content (e.g. text, images, video) by learning the patterns of an underlying dataset and then generating content that replicates those patterns as closely as possible in response to a user query.
Whether you are new to legal research or new to generative AI (or both!), there are specific limitations you should understand before you attempt to use genAI in this context.
Lawyers are bound by professional responsibility obligations. Any lawyer using new technology must do so only after ensuring they do not compromise these obligations.
►See Licensee use of generative artificial intelligence (2024) for the LSO's whitepaper on using generative AI.
In 2023, New York lawyer Steven Schwartz was sanctioned by a judge and ordered to pay a $5,000 fine for submitting a document with fake case citations generated by ChatGPT. The lawyer had not checked to ensure that the cases existed before submitting the document to the court. A number of similar cases have been reported by the media involving Canadian lawyers.
Mata v Avianca, Inc, 2023 WL 4114965 (SD NY).
Legal Research Online: Information Seeking in the Digital Environment by The most effective genAI tools for legal research are those designed and trained on legal information in the jurisdiction you are researching.
These Canadian tools are available either freely online or via the Law Library:
