In traditional litigation context, a typical document review process will involve human review of all documents to determine the relevancy and/or privilege (if any) attached to documents in the discovery process. However, with the ever-increasing number of electronic documents in the discovery process, artificial intelligence-based technology such as predictive coding is now a more feasible tool for review of large volume of documents in litigation.
Predictive coding is a document review process that entails the use of sample documents to train an algorithm to identify relevant and privileged documents in a large document set. The use of predictive coding allows for significant cost reduction in the document review process especially where the document size is extremely large.
The adoption of this seemingly new technology in litigation discovery process has not been without judicial controversy. It was the subject matter of intense judicial battles in many jurisdictions such as the United State, United Kingdom, Ireland, Australia etc. These judicial fights were based on whether a court can order parties to litigation to use predictive coding technology as opposed manual review to review documents for the purpose of complying with discovery obligations in civil litigation.
In the spring of 2012, in the case of Da Silva Moore, et al v. Publicis Group, the US Magistrate Andrew Peck for the first time approved the use of the technology in document review against the wish of the opposing party. The decision by Magistrate Peck ignited a trend of judicial approvals for predictive coding technology in other US courts as well as in subsequent cases in other jurisdictions e.g. Pyrrho Investments Limited v. MWB Property Limited and Brown v. BCA Trading Limited (UK), as well as Irish Bank Resolution Corporation Ltd & Ors. v. Quinn & Ors. (Ireland).
More than half a decade after the US courts judicially endorsed the use of predictive coding in document review, advocates of predictive coding technology in Canada were eagerly anticipating the first case in Canada were the courts would provide similar endorsement like those in other jurisdictions. That anticipation came to fruition in a recent decision on a motion before Canada’s Competition Tribunal in The Commissioner of Competition v. Live Nation Entertainment Inc. et. al. 2018 CACT 17. The Competition Tribunal is a specialized restrictive trade practices administrative tribunal composed of judges of the Federal Court and expert lay people.
Commissioner of Competition v. Live Nation Entertainment Inc.
The Commissioner of Competition (“Commissioner”) brought an application to the Competition Tribunal pursuant to section 74.1 of the Competition Act, RSC 1985, c C-34 alleging the Respondents engaged in deceptive marketing practices by promoting the sale of tickets to the public at prices that are not in fact attainable. In this motion, the Commissioner sought an order from the Tribunal compelling the Respondents to produce further and better affidavits of documents. Five of the eight Respondents had produced Affidavit of Documents (AODs) which listed no documents (“nil AODs”). In the nil AODs, the five Respondents stated “[t]he documents listed herein, if any, were located through the use of technology-assisted review…”. The Commissioner objected to the Respondent AOD on the ground that the search for the documents was clearly inadequate as it produced fewer documents than expected, document that the Commissioner had previously obtained from some particular respondent was not listed in the applicable AOD, several documents relating to the Respondents’ marketing practices, consumer behavior, and impact of the Respondents’ advertising were not produced.
Curiously, in their response, the Respondents sought to explain away the various deficiencies on the basis that they conducted searches in a more modern way by using computer assisted technology. The respondents’ document collection process had resulted in the collection of some 2.5 million documents. They used predictive coding technology to review the documents. Counsel reviewed some 8,287 sample documents (seed set) which was then used to train and validate the predictive model. The documents were then culled down to 55,000 documents, and completing the task within a relatively short period of time.
Unlike the US case of Da Silva Moore, et al v. Publicis Group, the use of predictive coding technology in the review of the documents in this Canadian case was not opposed by the requesting party – the Commissioner. The Tribunal noted that the Commissioner had not objected to the respondents’ use of predictive coding technology to assist in the review of the documents. The Tribunal then went further to provide what appears to be the most explicit endorsement of predictive coding technology by any Canadian court or tribunal. The Tribunal stated:
The Tribunal encourages the use of modern tools to assist in these document-heavy cases where they are as or more effective and efficient than the usual method of document collection and review. (at paragraph 15)
Before now, the most relevant case in Canada in which the issues of predictive coding arose (implicitly) was the Ontario Superior Court case of Bennett v. Bennett Environmental Inc. 2016 ONSC 503 (CanLII). Not surprisingly, the facts giving rise to the Ontario litigation originated in the United States. In that case, the plaintiff who was a former officer and director of the defendant corporation was extradited to the United States to face criminal prosecution by the United States Department of Justice for white-collar crime. Following his arraignment, the Department delivered a “document dump” to his US attorney. To review the documents, the US attorney used predictive coding for the first level of review to reduce the number of documents which was later subject to human review. The issue before the Ontario court relates to (among others) the reasonableness of the legal fees charged by the plaintiff’s US attorney for the review of large set of documents produced by the United States Department of Justice.
In his ruling, Mesbur J. upheld the reasonableness of the document review fees. According the Judge:
Given the use of predictive coding for the first level review of massive document disclosure, I do not find it unreasonable for the lawyer to then use paralegals to conduct the next level or levels of review. I make no adjustment on this account. (at paragraph 44).
The judicial approval of predictive coding technology in Commissioner of Competition v. Live Nation Entertainment Inc. went beyond the tacit endorsement in Bennett v. Bennett Environmental Inc.
Predictive coding technology remains the most efficient and cost-effective tool in review of large documents in civil litigation process. Predictive coding technology is here to stay. While the lack of judicial endorsements over the years has not stifled the use of the technology in Canada, it will be gratifying to see more Canadian judicial endorsements of a technology that has the capacity to revolutionize and modernize electronic document discovery process in civil litigation.