(I’m Still Not Here)

Hello Friends in the Computer,

I felt like taking a break from…my break…and thought I’d to pop in and drop some knowledge thoughts and update you on what I’ve been up to during my sabbatical. Although I don’t know if it’s technically possible to have a sabbatical from a side hustle, especially when the side hustle is something you do for free…

Also, I just drove from Northwest Indiana to southern Ohio by myself and had a lot of time to think.

So let’s start with the thing everyone is talking about and get that out of the way…

Generally speaking, my attitude towards this initial hype cycle of LLMs/Generative AI/ChatGPT remains this:

Which is not to say that I don’t find the technology impressive or that I don’t think there’s some real potential there. It is and I do. I just agree with what Sam writes here. My one change in his diagram is that instead if “profitable”, I would write “adopted.”

I’m still not 100% clear on what the magic ticket is to get a tool or technology adopted and integrated into a user’s or org’s workflow. I know how to really fuck that process up and how to make it as easy as possible, but I also know ROI, other benefits to self, and benefits to others are frequently ignored in any attempt to break the inertia and status quo of the legal world.

I’ve really been trying to figure out why I haven’t been all that hyped about the newest AI innovations. Everyone else is writing and speaking 1000s of words about it, why don’t I want to? Then I realized…the system is so broken that this will either just be a band-aid or hasten it’s implosion.

One of the things that tech evangelicals will say when you seem less than impressed with their solution is “Henry Ford used to say ‘If I asked people what they want they’d say ‘faster horses.’’” First of all, Henry Ford sucks on many levels and emulating him is not a goal of mine and shouldn’t be of yours. Secondly…any technology that is placed within the legal world, especially the A2J one, and doesn’t have collaborators that are also working to improve processes, data, access to information, and other inequities IS the faster horse.

I mean, there’s always value in trying to fix immediate issues, especially if you are relatively powerless to make the true systematic changes needed to really solve the problem, but if I may quote Taylor Swift, “Band-aids don’t fix bullet holes.” And double check to make sure you are actually solving a problem and not actually exasperating other issues that will cause a system held together with bubble gum and duct tape to finally crack.

(As an aside, I love when I’m snooping on Silicon Valley Twitter when things like DoNotPay hit the news and there’s some TechPerson [obnoxiousness knows no gender] with some bulllshit like “Working to democratize the law!!! ” in their bio and you go to look at their product and it’s something to help people with the paperwork involved in VC backed tech start ups and they want to believe lawyers are the problem and not the system that created the barriers that require a bar admission to solve. Anyway, bless their heart..)

Just to prove I’m not a hater, I will say this: Generative AI seems like less of a scam than NFTs and cryptocurrency and is not as much of a “solution in search of a problem” like other blockchain based applications. There. That was nice. Ish.

You heard about the lawyer that submitted a brief written with ChatGPT to federal court and it turns out that it was filled with citations to cases that don’t exist? I know you did, everyone has. I’m not finding a cite to link.

Okay, everyone gather around and take a knee….

Any public facing general generative AI tool, be it ChatGPT, Bing, whatever Google has - and yes even if you pay OpenAI $20 a month for the special version - is not an appropriate tool for creating most professional legal documents, especially if you’re not going to bother doing a review or independent verification of the content. Totally corncob move. And if you keep this up, in about 20 years or so, we’re going to get an even more strongly worded comment to a model rule that will suggest that lawyers not act like total corncobs when it comes to using tech and then 10 years later 80% of the states will adopt it and some of you will have to sit through another hour of a worthless CLE that teaches you nothing about how to not act like a corncob but does make some people in positions of authority feel like they’ve done something.

<breaks fourth wall and stares at the camera like a character in a mockumentary>

Again, I’m not saying don’t use Generative AI tools. I’m saying don’t be a corncob about it. Here’s some good resources:

• In this thread, Nikki Shaver compiles legal tech focused tools that are available for you to try.

• As Colin Lachance notes, his company JurisageAI has a tool that lets you check for fake cases. It’s free to use and covers Canadian law.

• Similarly, Jacqueline Schafer’s ClearbriefAI has a partnership with Fastcase and the Free Law Project that will check and flag fake cases.

<sigh> I really wanted to show you how I set up my gardens and make it a big allegory about legal tech but…this is too long already. Maybe another time.

Be well,

Sarah

p.s. Here’s the subscribe and share buttons so Substack doesn’t yell at me.

Thank you for reading Sarah's Commonplace Book. This post is public so feel free to share it.