Are Algorithms Required for Ethical Legal Research?

As we are increasingly aware, the ethical Duty of Technology Competence requires lawyers to keep abreast of “changes in the law and its practice, including the benefits and risks associated with relevant technology.” To date, 35 states have adopted the duty.

In a previous post, I highlighted the risks of blindly relying on algorithmic results (relevant technology) as a potential violation of the Duty of Technology Competence. We now have case law from Canada focusing on the benefits of using algorithmic results to perform legal research. In fact, this case law may be interpreted as requiring the use of algorithmic results when ethically performing legal research. 

In both Cass v. 1410088 Ontario Inc. (“Cass”) and Drummond v. The Cadillac Fairview Corp. Ltd. (“Drummond”) justices of the Ontario Superior Court made comments about artificial intelligence and legal research.

The Cass case was a slip and fall in which the defendant prevailed. The plaintiff, who was liable for costs, argued that defendant counsel fees were excessive and unnecessary. One issue raised was a $900 fee for case precedents, which the plaintiff argued, are available for free through CanLII or publicly accessible websites. Justice Whitten, perhaps also a lover of legal tech, agreed. He stated in relation to both the excessive amount of time counsel had spent on legal research, as well as the fee that, “[i]f artificial intelligence sources were employed, no doubt counsel’s preparation time would have been significantly reduced.” The defendant’s claims for disbursements was ultimately reduced from $24,300.67 to $11,404.08.

In Drummond, the defendant objected to the $1,323 claimed for legal research costs incurred using WestLaw. Justice Perell commented that the law is divided regarding whether a disbursement for legal research is a recoverable cost. One view is that legal research tools are simply part of a lawyer’s overhead and not recoverable, another is that they are a reasonable and recoverable disbursement.

Justice Perell’s own view aligns with the latter. In allowing the $1,323 disbursement for legal research he commented that, “computer-assisted legal research is a necessity for the contemporary practice of law and computer assisted legal research is here to stay with further advances in artificial intelligence to be anticipated and to be encouraged.” He further noted that, “computer assisted legal research provides a more comprehensive and more accurate answer to a legal question in shorter time than the conventional research methodologies.”

While the case law is from Canada, it provides insights into the foundation of modern legal practice. All lawyers perform some type of legal research when preparing a case. The major benefits of using algorithmic results (a form of artificial intelligence and a relevant technology used in law practice) for legal research is the efficiency and comprehensiveness associated with the results. If done properly, legal research performed on databases powered by algorithms takes much less time and is generally more comprehensive than performing research in print.

When using these Canadian cases to further interpret the Duty of Technology Competence, we see that a competent lawyer will understand the benefits of using algorithms for legal research. As such, attorneys should use these resources to ensure that legal research is performed comprehensively and efficiently.

On the flip side, while algorithms should absolutely be used for the utmost benefit, lawyers must still understand the risks associate with the use of algorithms for legal research and use reasonable care when relying on the results. While Justice Perrell in the Drummond case mentions that "computer assisted legal research provides a . . . more accurate answer to a legal question," that may not be true in every instance. Lawyers must not blindly rely on the algorithm to provide an accurate result. Competent lawyers must use their legal training and independent judgment to test and carefully review the results to determine any anomalies.


  1. In Cass, the judge indicates that the lawyer's arguments were contrary to basic legal knowledge and Supreme Court precedent. In other words, this was not a difficult legal research problem. I infer that the judge was saying that a reasonably competent lawyer would have found the relevant law quickly with an AI-assisted legal research system. But Cass could be distinguished in a harder case. Meanwhile, in Drummond, the judge indicates that computer/AI assisted research is recoverable if "properly done," i.e., if the expenditures of both lawyer time and computer time are "reasonable and appropriate for the particular legal problem." So both Cass (implicitly) and Drummond (explicitly) contain the notes of caution that you've raised about researching on databases with algorithms.

  2. It's wonderful that these judges are aware of the issues. Now to continue to sound the alarm to the rest of the legal field.

  3. I think there is a difference between background legal research on the matter as a whole and legal research to prepare a brief or trial. Those expenses should be specific on the bill on what the research was for without giving too much confidential information away. However, I think courts should be careful when looking at this issue. We want to encourage diligent and ethical work without creating an environment where every aspect of the lawyer's work will be questioned. That will not bring trust back to the profession.


Post a Comment

Popular posts from this blog

For The Love Of Archives

US News Scholarship Impact Issues

AALL/LexisNexis Call for Papers 2019-2020 Now Open!