Watch the full interview with Dan.
In this fireside chat, Dan Hunter, Executive Dean at King’s College London, sits down with Josef Co-founder and CEO Tom Dreyfus.
Together, they dig into why legal services are so ripe for disruption by GenAI, why more legal teams need to adopt the technology sooner rather than later and how businesses can begin experimenting quickly.
Watch Tom’s full chat with Dan to learn more, or check out our recap below.
“With transformer technologies and large language models, this is the end of the world as we know it,” Dan says. “There’s a new world emerging and I wouldn’t ordinarily have said that except for the nature of this technology.”
“The entire edifice of legal practice as we currently practice it is going to crumble in some way. Where, how, and how long it’s going to take is obviously the most interesting question,” Dan adds.
Because it deals in words.
“So much of law is the manipulation of words,” Dan says.
And while “lawyers are very proud and spend years and years honing their craft, there’s so much that is simply: take this, summarize it, put it into this chunk, then charge a lot of money for it.”
“That’s a lot of what legal practice involves and large language models can do that easily,” Dan says.
It just requires a little patience.
“Start by treating your tool as a paralegal, or an intern. You wouldn’t expect them to come back with a perfect answer if you just said, ‘Write me a brief’, right? Give them some context.”
“Explain what it is that you’re looking for, give your tool some examples of briefs that have worked in the past, and then ask it to do that. Tip it at the end or say, ‘Take a deep breath,'” Dan jokes.
“If GenAI's value curve is exponential where it starts to asymptotically become vertical and you miss it and wait six months, you can't catch up. You just can't. That's all of your money, all of your clients, all of the profits eaten up by your competitors who started before you.”– Dan Hunter, Executive Dean at The Dickson Poon School of Law, King's College London
“Sometimes a little bit of technical knowledge is good because all that a GenAI tool is doing is predicting the next most likely token, or word based upon what it’s seen in the past.”
Dan explains, “Once you understand some of those aspects, including word vectorization and why words work the way they do, then you can understand what each word is good and bad at.”
The pay off?
“You can ask your tool to do one thing and be confident that it’ll do a good job because you know how it’s doing, what it’s doing.”
Jump in now, is Dan’s advice.
“This thinking relies upon an assumption that the nature of the value curve on GenAI is exponential versus linear.”
“If it’s a linear curve, it doesn’t really matter when you adopt it. You can spend a bit more money later to catch up and learn from other people’s mistakes.”
“But, if the value curve is exponential where it starts to asymptotically become vertical and you miss it and wait six months, you can’t catch up. You just can’t. That’s all of your money, all of your clients, all of the profits eaten up by your competitors who started before you.”
“So when it comes to business strategy, the question is: Do you believe that LLMs have an exponential growth value profit curve, or do you think that they have a linear one? If you think they have a linear one, good luck. Don’t invest now. Just wait for people to screw it up. If you think that it’s exponential, then the only appropriate business strategy is to invest now because you have to get on board.”
How else are legaltech pioneers thinking about GenAI? Watch Tom’s chat with The Legaltech Fund’s Zach Posner to hear why he thinks the legal industry will “fundamentally change” in the next “three to four years”.
Book a demo to see first-hand how easy it is for your whole team to get up and running with legal automation.
We'll be in touch soon to arrange a time to speak.
