This semester we’ve talked a lot about the legal field’s resistance to change. I, like the majority of the class, understand the need for legal innovation. Working at a firm this summer, I realized even more the necessity of legal innovation. For example, doing a 5000-e-mail doc review manually showed how useful and cost-efficient document review technology is. However, I understand the hesitancy about automation as well.
This summer, I was at work scrolling through LinkedIn when I saw an ad for Alexsei. I clicked on the website, and for some reason, I was flooded with panic.
Alexsei is an online legal platform that provides answers to legal questions in memo format. To use Alexsei, you submit a legal question with facts to give context and then A.I. and research lawyers “review and synthesize millions of documents to find the most relevant case law and legislation.”[1] Through this review, Alexsei produces a research memo which includes a conclusion, caselaw summaries and a full list of authorities. The part that spooked me most was reading that using Alexsei for legal research saved about “four hours of work per research task.”[2]
I remember feeling genuinely worried that my firm would discover this website and realize that they don’t need summer students. Thankfully, my rational brain came back a couple of minutes later, and I remembered that the sole purpose of a summer student isn’t to do legal research. One thing I wondered was how A.I. could assess some of the more nuanced / context-specific parts of a legal question.
This same question has come up during my team’s efforts to build a legal advisor app that, to a certain extent, evaluates the strength of a human rights complaint. For context – a human rights complaint has a three-part test – 1. that you have a protected characteristic, 2. That you were treated negatively in a protected area, and 3. Your protected characteristic is at least a factor in why you were treated negatively (a nexus between the discrimination and the protected characteristic). When we came up with this idea, our team’s biggest concern was how do we automate the third question. The crux of a human rights complaint is that there is a nexus. So how do we ask a series of plain language questions that allow us to assess the strength of a claim? Obviously, in the strongest cases, the user/complainant will have some evidence. They were fired the day after they told their boss they were pregnant. An employer fired an employee and said it was because of a protected characteristic. An employer ignores an employee’s formal complaints of racism coming from other employees. However, the tricky thing about discrimination is that often times its subtle.
The other day my team had an opportunity to sit down and speak with Laura Track – the director of the B.C. Human Rights Clinic. While she was super excited by our app, she brought up our exact worries. She wondered how a series of questions could accurately capture the subtlety of some discrimination. If our app doesn’t catch that subtly, are we deterring complainants with good claims from proceeding to the forum?
This is still an issue we’re working through, but I think my worry with our app extends to all A.I. technology that offers a legal opinion. E-discovery and doc review technology obviously doesn’t have that problem to the same degree, and I really do see the value in automation in a lot of areas, but I can’t seem to wrap my head around an A.I. like Alexsei being able to come to a legal conclusion on discrete and subtle legal issues. Maybe I don’t fully understand the capabilities of A.I. (a very likely possibility), and maybe the A.I. just conducts preliminary research that the “Research Lawyers” mentioned on Alexei’s website review and finalize.
When I looked into more, though, I realized how Alexsei was used in practice. In an article written by Canadian Lawyer Magazine[3], a Saskatchewan-based lawyer explained how Alexsei saves her time. She said that in situations where she needs to research very thoroughly and find specific (what I assume means analogous) cases, she doesn’t stop at the Alexsei memo – she uses Alexsei to save herself “two or three hours” of her own research by getting her the leading cases in the area. One quote stood out to me specifically “I think it makes us better lawyers because we’re on top of things, and that energy can be better spent doing what we’re intended to do – share our experience and advice with the client.” The lawyer even goes on to say that it helps her articling students because they can go on to “finer search” after Alexsei does the initial review of caselaw and legislation.
[1] https://www.alexsei.com/testimonials
[2] https://www.canadianlawyermag.com/practice-areas/personal-injury/how-alexsei-streamlines-firms-legal-research-needs-through-ai/363702
[3] https://www.canadianlawyermag.com/practice-areas/family/a-tale-of-two-firms-an-alexsei-case-study/365083
This is such an interesting topic. I am very intrigued by Alexsei and I’ve talked to lawyers who have found it very helpful for making their research faster and more efficient. It’s been said that technology won’t replace lawyers but will work with us; the lawyering of the future will be done in collaboration with technological tools that speed up the repetitive parts. And the lawyers who are proficient at ‘collaborating’ with technological supports are going to have a competitive advantage.
Writing this blog post really helped me shift my perspective to the one you’re describing in your comment! I think back to my summer and how many hours I spent scouring all the different databases to make sure I didn’t miss any authoritative cases. With a program like Alexsei – I could have spent those hours doing more pointed research.
Your last point is also a great point that I hadn’t fully considered!
Great blog post! After reading your post, I searched Alexsei on google to see what kind of tasks the application assists users with. I saw they had the tagline “Effortlessly turn any research question into a comprehensive memo.” I wonder if the application can deal with multi-faceted legal questions that are more complex. I also wonder whether Alexsei provides an opinion memo on each issue, or if it is more of a summary of the law.
This is so interesting, and it actually reminds me a lot of how I did my research in grad school. For my thesis, I did a qualitative analysis of newspaper articles with the help of some qualitative research software. It definitely was not AI-based, but it did save me many, many hours. Similarly, though, at the end of the day, there still needs to be a human on the other end for the more nuanced themes/coding.