Background and the dispute
There have always been disagreements about the scope and sufficiency of discovery efforts where litigants are dealing with large-scale documentary discovery. The issue has become even more fraught where new AI technologies are deployed in attempts to make the process of identifying relevant documents more efficient. The recent decision of the BC trial court in Acciona Wastewater Solutions LP v Greater Vancouver Sewerage and Drainage District, 2025 BCSC 1256 (CanLII), explores some of the issues that can arise, and what standards are to be used by courts in determining what constitutes judicious (or improper) use of AI in gathering and producing documents to another party.
In this decision, the defendant, Greater Vancouver Sewerage and Drainage District ("GVSDD"), brought an application for an order requiring the plaintiff, Acciona Wastewater Solutions LP ("Acciona") to amend its productions amounting to 4 million documents, claiming they had produced many irrelevant documents and their process amounted to a data dump. For its part, GVSDD had produced only 225,000 documents.
Competing AI review methods
Acciona leveraged a technology-assisted review process known as TAR 1.0 to review their documents. GVSDD leveraged a Continuous Active Learning process (CAL) also known as TAR 2.0. GVSDD argued that Acciona did not ensure the AI model was adequately trained to identify relevant documents, and that recall rates were lower than industry standard.
The court's decision
The court found that Acciona's production was not a data dump, and that it had complied with a comprehensive document production protocol and agreed upon search terms with no evidence that it acted with the intent of obfuscating the discovery process or offloading its discovery costs onto GVSDD. The court also found that Acciona should not have to amend its list of documents, as the rules of procedure should not apply rigidly in cases involving very large numbers of documents. To establish that a list of documents is "inaccurate," there should be evidence showing a "dilatory and casual attitude to production."
The court agreed with the defendant that "meaningful, reliable, and complete" production should allow the receiving party to review the disclosure without the added burden of excluding an undue number of irrelevant or duplicative documents," but "the standard is not perfection." The court also referred to the Sedona Canada Principles emphasizing the overarching importance of common sense and proportionality to address increasingly high volumes of electronically stored information and the high costs of litigation.
The court dismissed GVSDD's application and awarded Acciona its costs of the application in the cause.
Read the original article on GowlingWLG.com
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
