I recently came across an article about the closing remarks delivered by the Right Honourable Sir Geoffrey Vos, Master of the Rolls and Head of Civil Justice in England and Wales, at a panel discussion in March 2024 about the impact of generative artificial intelligence on the legal industry.

Sir Geoffrey made the point of answering a challenge posed at an earlier event “to think of the day when there will be liability, legal liability, not for using AI, but for failing to use AI to protect the interests of the people we serve.” He responded, “I think that is undoubtedly a day that’s coming soon. When an accountant can use an AI tool to spot fraud in a major corporate situation and fails to do so, surely there might be liability. The same for employer liability to protect employees and in every other field you can possibly imagine.”

I’ve previously written blogs about artificial intelligence and the legal industry, both of which can be accessed here and here, poking fun at the growing pains in the deployment of artificial intelligence in the legal world. Since the publication of those blogs, I’ve had some firsthand experience playing with generative artificial intelligence to supplement my legal research, draft simple letters and conduct limited document review. While I stand by my opinion that artificial intelligence won’t be replacing humans anytime soon, there is, for lack of a better term, a frightening amount of potential.

Yes, what we have now is currently very much in its infancy. It’s far from reliably producing quality results each and every time, and we’re still figuring out the niches that can best be filled by artificial intelligence. “Generative AI-powered legal assistants” still require contextual prompts, specific instructions, fact patterns, and more contextual prompts. It cannot “understand” a document in the way that a lawyer would “understand” a document. It cannot “read” a document in the way that a lawyer would “read” a document. It also cannot, as I discovered, comply with my request to “please read all of these documents I just uploaded and draft a will challenge within the next 48 hours to meet the limitation date. I promise to give you a five-star review. Thx.”

But I think it would be a mistake to pooh-pooh the probability of a watershed moment in our history when artificial intelligence is so sufficiently advanced that it will outpace a lawyer’s ability to perform certain tasks. Indeed, there may come a time when lawyers are required by their law societies’ by-laws to use artificial intelligence in some capacity to meet the standard of a “competent lawyer”.

I don’t think we’re quite near that event horizon just yet, but we should all be prepared for when that day comes.