A Few Thoughts on the Future of Law

Had an interesting Twittersation last night with Dr. Sara Kubik, a Ph.D. in Technology (and current third-year law student.) First I Storified it, which I’d never done before, so here’s that:


But I thought I’d expand on some of my points a bit.

1) Losing a lawsuit is the only thing worse than winning a lawsuit.

Lawsuits suck. They purely suck. It doesn’t matter how good your case is: every lawsuit I’ve ever brought, I brought because someone had out-and-out stolen from my client. No fair use defense, no inadvertent infringement defense. Likewise (I am really good at picking clients) every time one of my clients has been sued, the claim could at best have been described as “colorable.” That’s a legal term of art which means “I probably can’t report opposing counsel to the disciplinary commission for filing frivolous lawsuits.”

And I’ve won every time.

(Well okay we settled one but it was pretty favorable.)

My clients still had to go through a very distracting, draining legal process which cost them a lot of money up front. While we recovered costs of suit in most cases, sadly I’m not as good at picking lottery tickets so we never got a zillion-dollar payday. Maybe that would make a suit worth it, but frankly given the level of disruption they cause a business that’s what it would take, if I cared about money. I had to file most of those suits because there were intellectual property protection concerns. In that respect, I won – my clients’ IP was protected. But I wish I could have found some other way.

2) Rocket Lawyer/Legal Zoom/Et Al and their impact on the legal profession.


Okay, I can unpack that a little. I have seen people get work from “inexpensive legal services providers,” (hereafter, “ILSP”) without naming names (and not necessarily either of the two providers given as examples) which required extensive additional legal work to… well, to fix, to be blunt. I know other lawyers have had similar experiences.

Without knowing how much of what kind of legal work ILSP provide, I can’t give accurate evaluations of whether they are cost-effective in the long run for individual clients and/or society. I admit that freely. But the numbers I used above in the Twittersation seem, if not provable, plausible. It only takes one actual court case or even the preparation for one to wipe out the savings from a very large number of inexpensive legal services.

The people who work for ILSP are, one fervently hopes, licensed and competent attorneys. But given the nature of such operations, they simply can’t become as familiar with a particular client’s needs and issues as a personal counsel can. The numbers don’t work. That extra hour to review, that extra bit of specialized knowledge… they are not worth saving a hundred bucks over. In legal services, as in few other things, you get what you pay for.

3) AI and the future of legal work.

My reference to “Happy Fun Space” is a shout-out to Iain M. Banks’ “Culture” novels, where we are told some AI become so computationally powerful that they spend their time in simulated universes of their own design, which they call”Infinite Fun Space.” The “Singularity” is a hypothetical point in history where human beings or their AI colleagues become so powerful that they can transition mankind to a post-scarcity society. In short, once we get AI powerful enough to do what I do, they’ll be so powerful we won’t have to do anything at all. (This could go really wrong, or really right, depending on whether we get AI of the type Scott Alexander refers to as “Moloch” or “Elua” in his amazing essay Meditations on Moloch.)


However, even absent Skynet or Mother, AI will still have a significant impact on the legal profession as technology advances. There are already AI (though they aren’t true AI, they’re agents) which can draft simple contracts or help people protest parking tickets. For parking tickets, this is probably okay. But even for simple contracts, I am not convinced this is a good idea at all. I am fond of saying that an AI could probably safely do a big chunk of my job… but part of the chunk it could not safely do is accurately determining which chunk it could safely do. Any transaction more complicated than buying a pack of gum at the drugstore is inherently individual. AI aren’t there yet. Human programming isn’t there yet. Subjective judgment and experience are holistic and intuitive and that’s just how it is.

For instance, say you’re an attorney for a private, closely-held company which is subject to significant regulatory restrictions. (Yes, this is an example from experience.) You want to redraft the bylaws for your client’s corporate entity. Here are just two things an experienced lawyer will think about which have nothing to do with the actual bylaws.

a) You only have a handful of shareholders. What are their estate arrangements? If one of them passes away or becomes incompetent, will their ownership interest pass to an individual or individuals who will need findings of suitability? Or worse, who can’t get findings of suitability? What will that mean to your client’s regulatory compliance?

b) Maybe you’re lucky and your client’s ownership structure is such that it qualifies for WBE or MBE status. (Don’t know what that is? Does the AI?) First, analogous question to the above. Second, are you sure the bylaws structure control such that in the event of a shareholder conflict, you always have the outcome required to satisfy the WBE/MBE certification agency?

Did whoever programmed your “Makin’ Corporations!” AI/agent think about things like that? Did they KNOW about things like that? Do they keep the program constantly updated with legal, business, and tax strategic background? Can you sue them for malpractice and report them to the bar if they don’t? This is the kind of thing I mean when I say that 999 times in a thousand, the client may save a small amount of money, and the thousandth time, they may end up bankrupt, in jail, or both.

Anyway, those are my thoughts on the matter. Comments invited.


Leave a Reply

Your email address will not be published. Required fields are marked *