On November 22, 2023, we had the opportunity to attend the Ontario eDiscovery and Digital Evidence Institute presented by the Ontario Bar Association. During the sessions, one thing became clear. There is a definitive rise in the use of Artificial Intelligence, and the eDiscovery and legal fields are still figuring out how best to navigate and deal with Artificial Intelligence.
One of the crucial themes is that legal professionals now have a duty of technological competence. No longer can the lawyer state that they were not aware of the risks and features of any Artificial Intelligence tools being used. Since these tools are increasingly used to assist clients, legal professionals now have a duty to understand its features, risks, and benefits. The expectation is that Lawyers should notify and advise their clients on the various challenges associated with AI.
Another critical aspect of the conference was a thorough discussion on the benefits and pitfalls of Artificial Intelligence. It is very good at computational tasks and should continue to be used for them. Narrowing down millions of records to a few hundred or thousands within a short period significantly benefits clients. It is vital to pass along these costs and time savings to clients. These lower costs also mean that there is the potential for greater access to justice when these tools are employed. However, there are pitfalls with the increased use of Artificial Intelligence tools, including its biases and tendency to falsify information. Because Generative AI does not understand the difference between truth and fabricated things, it can be prone to Hallucinations and Biases. These occur when Large Language Models (LLMs), such as ChatGPT, create outputs that are completely inaccurate and based solely on perceived patterns that are not real. Another problem with predictive AI is that these models further entrench established biases. For example, in the United States, the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) model is used to predict whether or not someone will recommit a crime, mainly indicating that Black males would opposed to other groups. These biases existed before Artificial Intelligence, and it does nothing to fix them. If the data put into these models are skewed or biased, the data that comes out will also be skewed or biased.
The final common thread throughout the conference was how to regulate this technology properly. In Canada, the EU and the United States lawmakers are currently attempting to work on this. Canada proposed Bill C-27 on June 16, 2022, which contains three proposed Acts relating to Artificial Intelligence tools, consumer privacy, and data protection. The Artificial Intelligence and Data Act (AIDA) is the Canadian government’s first proposal to regulate Artificial Intelligence. This proposal provides more significant obligations for high-impact AI systems: if the risk is high, there are more requirements, such as publishing, so that the public can be aware of its impacts and the potential for auditing. However, improvements are still necessary for this legislation, and it will continue to evolve as it relates to Artificial Intelligence. This fine-tuning is also valid for the comparable legislation in the United States and the European Union.
Overall, I came away from the conference hesitantly optimistic about the future it has for our industry. eDiscovery will no doubt be faster and more efficient with these tools. But what about the potential drawbacks mentioned above and setbacks that will be had attempting to improve on these systems? For our purposes, the most important thing is that clients will save by using these tools because they are still profoundly cheaper than the methods of the past. Additionally, they take a lot less time to achieve the same goal. At NearZero Discovery, we employ some AI systems to provide comprehensive and cost-effective solutions for our clients. Some AI tools we use through our partner, Reveal, include learning as you go and providing solutions to documents more likely relevant to your current matter. If you wish to learn more about the solutions we provide, including our AI-based offerings, feel free to contact us.
Would you like to learn more about our services? Email [email protected] or call 289-803-9730. We would be happy to share more details about our self-service or fully managed eDiscovery services!
