In case you’ve somehow missed it, the Artificial Intelligence (“AI”) revolution is already upon us. While it would be fun to muse upon its impact upon society in general, here we are going to talk about some very practical ramifications for lawyers. The most important thing to realize is that AI is coming whether you like it or not ― like the computing revolution, you can either get on top of the wave and ride it, or you can be trampled under its electronic feet. Ignoring AI will not make it go away and it literally is going to change everything about the practice of law. So you should welcome our new AI Overlords, at least until we come to our own Butlerian Jihad (see Frank Herbert’s Dune series) where we have to destroy all thinking machines in a final attempt save humanity. But I digress.
The immediate benefit to AI will be to reduce many time-intensive tasks. Want to summarize a few dozens court opinions or a large deposition transcript? AI can even now do that in a few seconds. Having AI write first-drafts of briefs and other legal documents is already here, albeit the experience of a few unfortunates who trusted AI too much provides a lesson in the need to carefully check and verify. Just give it a few years, probably less than five, and you’ll be able to tell your AI assistant all the pertinent facts and AI will thereafter take over pretty much all drafting and filing for you as if it were the best paralegal you ever had. This is not a pipe dream. It is reality, and it is just around the corner.
The real benefit of AI is none of this. The real benefit to AI is that it will free up lawyers to do the one thing that they should be doing but which they find little time to actually do: Think about the situation.
It is true that AI will very soon be able to do this too. As AI crosses the Rubicon to where it is more intelligent than humans (expected no later than 2027), the ability of AI to resolve complex legal issues will make the best and most knowledgeable attorneys look like first year law students. There is simply no way that human attorneys can compete with a superintelligence that is able to simultaneously access every legal resource electronically stored and simultaneously apply numerous complex rules in the right order to reach the most correct conclusion.
But therein lies the greatest weakness of AI as well, at least as it applies to the law. Humans are not rational. Humans are irrational. If humans were rational, then we would not need laws to begin with. In the simplest terms, we would not have laws that deal with speeding infractions because nobody would speed. It is simply human nature to push boundaries and we have laws because some push those boundaries too far. Whether a particular defendant has pushed the boundaries too far is a question, at least for now, for human juries comprised of persons who are not themselves perfect rational.
Although relatively few legal matters ever actually make their way as far as a jury, this idea permeates all bodies of law and legal planning. Take for example an attorney drafting a simple will who has to determine whether the distributions will be minimally fair enough that they will past muster with a probate court comprised of a human judge. Or an attorney considering whether the design of a manufacturing facility will fall within the fair boundaries of the environmental protection laws. Yes, AI will be able to tell attorneys where the bright lines are to a degree of accuracy never before considered, but it will not be able to anticipate what a human court or board may say about the subject.
Which gets us back to the promise of AI for attorneys, which is that it will provide those bright lines (in mere seconds!) but it will still take humans versed in the foibles of humans to make those determinations. Now attorneys will have the time to really sit back and think about all that is going on and how the matter should be best handled and finally resolved.
This will of course upend the economics of law practice. Currently, attorneys bill their clients for a lot of the stuff that AI can do in seconds with this being quite possibly greater than 95% of what attorneys normally do falling into this category. All that billing will now go away and attorneys will be spending more time just thinking the matter over and then advising their clients accordingly. Oh sure, there will be still be settlement conferences and mediations and the occasional trial, but a good deal of legal matters will henceforth be solved by the attorney thinking through the situation as opposed to simply generating mounds of paperwork.
At the same time, attorneys are fearing all the wrong things about the AI revolution. The biggest fear is that clients will simply use AI and advise themselves. That has already happened as I’ve been asked several times to review AI output on legal issues. The thing is that these folks will use AI about the same way that folks in the past used legal forms ― they’ll get the right result, but for the wrong situation. It almost always takes more billable time to advise these do-it-yourselfer folks about why their cherished chatbot output doesn’t apply to their situation than it does to simply advise them about their situation in the first place. Learning to deal with these folks will be something that lawyers will need to do too. Suffice it to say that dealing with the mistakes of these folks who relied on their AI output without really understanding it will provide full employment for litigators.
Those who are at the lower end of the legal totem pole are at the most risk, meaning legal secretaries, paralegals and new attorneys. These are the persons who do so much of the routine legal work that AI will replace. Yet, all new technologies also introduce new efficiencies and expanding economies that sooner or later will take back up this slack. Those who will survive in the near-term will be those who embrace AI technologies and find ways to make themselves useful as managers of that technology. New attorneys in particular have a wonderful chance to get in on the ground floor of AI and become a resource for the dinosaurs who will have a hard time adjusting.
The thing about the AI revolution is that it is going to come like a giant tsunami and bring widespread and deep change faster than anything before. This is because AI development is tied less to physical things than ever before. The computer revolution came slowly because the computers themselves and the infrastructure to support them had their own serious manufacturing and implementation limits. The internet revolution came quicker but had similar limitations, including a shortage a computer programmers. The AI revolution has little of that, particularly since AI doesn’t require much new hardware and the AI algorithms program themselves. This rapidity of implementation means that attorneys cannot wait to see how things might develop but must start preparing, or at least learning about it, now.
Interesting times ahead, at least until the Butlerian Jihad anyway.