Bankruptcies, AI’s use in insolvency process on rise, says Davies report

By Christopher Guly

Law360 Canada (November 6, 2023, 9:37 AM EST) -- The first half of 2023 featured the highest number of business bankruptcies in Canada since 2019, with the country’s retail trade sector particularly hard hit by “uncertain global and Canadian financial conditions,” according to a report recently released by the financial restructuring and insolvency group at law firm, Davies Ward Phillips & Vineberg LLP.

It noted that the Office of the Superintendent of Bankruptcy reported 2,160 business bankruptcies and consumer proposals between January and June of this year, the highest number of business insolvencies recorded in the first half of a calendar year since 2019 – a 78.4 per cent increase from the lowest reporting year (2021) over the last five years.

“This trend lends support to claims that the Canadian economy is observing a rolling recession,” said Davies Toronto-based partners Natasha MacParland and Robin Schwill in the study that they authored.

“Although this overall increase is dramatic,” they wrote, it “should be viewed in the context of historically low levels of insolvency filings throughout the pandemic.”

The online paper said the number of proceedings under the federal Companies’ Creditors Arrangement Act (CCAA) reached a historic high of 31 in Canada during the first half of 2023, surpassing the 27 proceedings recorded during the same period three years ago.

In the first two quarters of this year, the largest increase in the number of insolvencies occurred in four sectors: retail; accommodation and food service; professional, scientific and technical services; and manufacturing, which accounted for nearly half (48.4 per cent) of all CCAA proceedings with 15 filed.

The Davies report also looked at how programs, powered by artificial intelligence (AI), “are becoming increasingly important practical tools for insolvency practitioners.”

“AI can be a game changer for data capture, process automation, error detection and real-time document processing,” said the paper, which noted that “for lawyers, creditors and stakeholders as well as regulators and courts that contemplate adopting AI tools, there is significant potential to improve work efficiency and quality.”

MacParland told Law360 Canada that there is a role for AI in claims administration and adjudication.

“For a larger filing where there is a lot of people filing claims, we see a role for AI there and potentially for some online dispute resolution,” she explained.

She and Schwill expect AI to play a role in e-discovery, which they said has “transformed litigation practice more generally by using predictive coding to classify electronic documents” and with AI will make it “possible to reduce the human element of the review process.”

“However, given the high level of discretion and sophisticated expertise required in this area, counsel must still provide the specific advice that debtors, creditors and stakeholders need,” said their report. “Decision-making and advice in this context can certainly be improved through the use of e-discovery technology that identifies patterns and concepts, and offers visualization tools that support litigation strategies.”

There is also a potential role for AI in predicting credit default.

“Because bankruptcy is a rare occurrence, to this point machine-learning has had insufficient information to understand the sets of attributes that can lead to a future bankruptcy,” said the Davies report. “Attempts are currently underway to overcome this challenge by combining the most successful individual algorithms to achieve a higher degree of forecasting accuracy. At the same time, given the discretionary elements involved, it is unlikely that AI will be able to determine the decisions that companies will make on timing of filing and the appropriate procedure on the basis of a complete analysis of the current state of the law.”

The paper points out that ROSS Intelligence, the first AI research tool used by a law firm, “was in fact trained with specific insolvency expertise. ROSS expanded to cover other areas and was similar to earlier research tools such as Westlaw, but with less required of the lawyer to ask questions. That is, lawyers would type in questions using natural language to find relevant up-to-date case law or to extract quotes or precedents in support of an argument.”

“However, ROSS Intelligence ceased operations in 2020 due to ongoing litigation with Thomson Reuters,” the report noted. “As any lawyer who has used legal research tools knows, no technology has to this point been able to replace the strategic thinking that goes into negotiating a complex restructuring or sale process, let alone ask the questions that the lawyer should be asking tools like ROSS. Important human dynamics are involved behind the commercial decisions that are made, and a deep understanding of more than the law on the books is necessary for any transaction. At the same time, lawyers are now expected to automate, to the extent possible, parts of legal research and the analysis process; they are also expected to make informed decisions about which parts can be automated.”

MacParland said that AI could replace “administrative bodies” in the insolvency process, but “not the actual strategic bodies.”

“Part of the value that lawyers bring is that they know the law, but they also understand the practical realities — they take what they know about the client and the client’s business and the industry they’re in, and use legal tools to help the client achieve their objectives,” she explained.

However, in her report, MacParland noted that “AI technologies can help improve access to justice” for third parties and stakeholders “by facilitating access to legal knowledge.”

“While AI cannot replace lawyers and does not remove the need for nuanced legal advice, AI tools can help explain legal jargon and processes in easy-to-understand terms by using natural language processing capabilities,” said the paper.

“Generative AI can even offer information that can serve as a starting point for stakeholders in their own legal research. In other words, AI can serve as a source of power to those who do not have legal training by serving as a knowledge resource.”

It can also, as the Davies report offered, have a place in the alternative dispute resolution process “in the form of mediation and arbitration [that] has been playing in Canadian bankruptcy and restructuring cases.”

“In other areas of law — for example the British Columbia Civil Resolution Tribunal — we have seen the growth of online dispute resolution (ODR). Technological innovations enabling online claim diagnosis, negotiation and case management — all contribute to the popularity of ODR,” according to the paper.

“AI tools in this space allow less sophisticated claimants to strategize their case or pursue remedies in cost-efficient ways as well as improve access to justice for claimants in remote areas or with limited resources. If adequate safeguards to protect privacy and non-biased decision-making as well as increased information accuracy can be built, there is great potential to use ODR for claims resolution in the insolvency context.”

And AI is appearing in the courts, which “have begun to issue practice directions and other policies relating to the use of AI,” said the Davies report. “On the heels of a New York case in which lawyers submitted fictitious research generated by ChatGPT in their claim, some Canadian courts have expressed their own thoughts on the use of AI.”

In June, the Court of King’s Bench of Manitoba released a practice direction regarding the use of artificial intelligence in court submissions. In the direction, Chief Justice Glenn Joyal said that to address the “legitimate concerns about the reliability and accuracy of the information generated from the use of artificial intelligence,” the use of AI “in the preparation of materials filed with the court” must be identified.

Also in June, the Yukon Supreme Court issued a similar practice direction.

“While the directions released by Canadian courts do little more than double-down on the existing provincial rules of professional conduct, they should be seen as warnings to the legal community,” said the Davies report. “The commonality across all jurisdictions is that courts are recognizing AI will be used in proceedings in the future but are attempting to grapple with the policy concerns regarding its dependability.”

“Judges are making clear that AI is still in its infancy, and the human component of law cannot be completely erased — at least not without disclosing this fact to the courts.”