Queen's Law held a panel discussion last week on the topic of AI and law. A recording of the discussion and a summary will be made available in the fall. In the meantime, here are my prepared remarks. The actual discussion ended up covering different ground than the questions that were distributed ahead of time and my fellow panelists had a lot to add. I'll post a link to the discussion when Queen's does.
1. How is AI already being used in the legal profession?
Frankly there's nothing that I would call AI that's being used in the legal profession (when I think "AI" I think "strong AI"). The closest thing to it is e-discovery software that can identify connections between people (some of which claim to look for things like wrongdoing) and potentially "hot" documents. There's also software that can extract sections of contracts such as Kira Systems (formerly Diligence Engine). IBM has been making claims about their progress in this area but I haven't seen it.
2. How do you think AI will be used in the legal profession over the next 10 years?
I expect the main thing to improve will be how legal research is done. Computer programs should be doing the first pass. The volume of material that lawyers are expected to know about it is too vast for even the secondary material + follow-up research with software to work (at a reasonable cost). Legal tech can help bring down these costs and help lawyers give better answers faster. Will it be AI? Probably not. But it might look like AI.
3. How do you think an IBM Watson-powered lawyer, in particular, will have an impact on the legal profession?
I've written a critical blog post on this topic that attracted a fair bit of positive attention. But even if we assume that Watson ever works as IBM hopes it will, I think the main impacts will be on legal research and making more law available to more people. The current legal database search tools work pretty well but they haven't changed in 15 years. There's a huge potential for making legal research better and opening up legal research to the unrepresented litigants who are an ever greater percentage of cases before the courts.
But the beauty of technological development is that it comes from very unexpected places. Ten years ago we would have been talking about how Blackberrys will evolve but now we're wondering where Android is headed and large-format, super high-resolution touch screens have taken over the market. Twenty years ago we could have discussed how pagers will evolve. Progress happens in unexpected ways.
Here's a Fordham Law Review paper from last year on the coming "great disruption": http://ir.lawnet.fordham.edu/cgi/viewcontent.cgi?article=5007&context=flr.
4. Do you think AI will terminate jobs for lawyers, create new ones or both? Why?
A powerful AI would eliminate many lawyer jobs. The more powerful, the more jobs go. But if there is a powerful enough AI that it's displacing lawyers then there'll have been a huge shift in the economy. The massive boost to productivity should improve the welfare of lawyers overall but at that point we'd have to start questioning what it means to be a lawyer. Perhaps the traditional lawyer will disappear and a new job will appear that we don't have a name for yet. Almost everyone knows what a car mechanic does but how many people have ever heard of a farrier? The job description of lawyer has been remarkably unchanged over the last 100 years but I'd be surprised if the next 100 doesn't see it breakdown.
5. What opportunities do you foresee for junior lawyers in a profession that could be revolutionized by artificial intelligence?
The tech-savvy lawyer is not actually very valuable now (look at who the country's richest lawyers are) but if AI revolutionizes law then there'll be a big demand for programmers who understand the law. I think the opportunity would be in getting the AI to do something different.
For people who have no technical know-how I assume there will always be a place for them with complicated bespoke advice (built on AI work).
6. How can law students and junior lawyers be taught to harness new technology in order to best serve them in their careers?
With an introductory programming class.
Not so that they can become professional programmers but so they can understand what programmers and computers can do for them. If you don't know what tools are available then you're not going to be able to use them.
7. Do you agree with Stephen Hawking and Elon Musk that AI could someday overtake humans?
Yes. We've been on a steady upward curve since the wheel. Humans get better every day at harnessing technology and calculating more things faster. Unless there's some sort of cataclysmic setback in progress we will eventually produce a computer that exceeds human capacities on all the relevant metrics. I'm not sure how it could be otherwise.
In the meantime, lawyers are still using the same word-processing tools they were using when NWA was rapping about pagers. We as a profession need to start adopting the tools and processes that have boosted productivity in the broader economy.