Categories
Legal ethics Off-Topic

Several scattered thoughts about advancements in AI

So, life really moves fast. Though, while it seems like advancements in AI are also really moving fast, maybe they aren’t moving as fast as the hype.

That is the topic for today. Now, when I say life really moves fast … I mean that between my starting this post yesterday and today a giant piece of what needs to be said has changed. It involves some of the ongoing saga of the DoNotPay tool and even more specifically raises questions about just how useful the tool maybe ever has been. But we’ll leave that discussion for the end.

First, ChatGPT. Surely if you’ve found your way here you’ve read something about it by now. The most recent high-impact news is of a massive investment by Microsoft in OpenAI the company behind the product. But also, there has been a good amount of discussion all over the place about whether the things this can do are good or bad for lawyers. (Usually, lawyers are the ones asking that question that way since lawyers often tend to think in that very insular fashion as opposed to thinking more expansively about whether it is good or bad for other people.)

The stuff I don’t find very compelling at all are the various claims of what sort of licensing tests ChatGPT can or can’t pass. I feel that way because (1) we are also still having productive discussions as a profession about whether standardized tests such as the bar exam has ever really measured what makes for a good lawyer; and (2) I’m pretty sure that, if you gave me access to all of the information in human history, I could also pass any test you put in front of me.

Along those lines, one of the most useful pieces I’ve seen in the last month or so is an article about a variety of great ways that General Counsels could make sure of ChatGPT right now. Of course, despite almost completely remembering the title of the article, my many attempts at Google search has left me unable to find it. One part of it I remember vividly was a part that talked about how a prompt could be used to rewrite a message into a more user friendly syntax when turning something “legal” into relatable to non-legal employees. If any of my readers can find the link to the article to prove I’m not hallucinatng, please post it in the comments.

Edit: Article found. But by me so this way I don’t have to get clear proof that I have no readers. The key to finding it was belatedly remembering that “use cases” was in the title. Linked in 5 use cases article

It is a good article not only because of its practical information but because it focuses on a mindset of how lawyers can look at these kinds of developments in technology as a way to better at lawyering rather than as a threat to whether people will need lawyers at all.

Now, let’s turn back to DoNotPay which I wrote about extremely recently. I was taking issue with something that it was trying to do that was a real transformation from what I understood to be their existing attempts at solving problems for people in need of legal services. The Twitter thread below that is worth reading from start to finish raised real questions for me about whether I was giving too much credit to what DoNotPay was otherwise offering.

Kathryn Tewson on Twitter: “I’ve been going in pretty hard on @DoNotPay and @jbrowder1 for the past couple of days, and I’ve had a lot of people defending the service, saying that it could be a real boon to those who can’t otherwise afford legal aid.” / Twitter

And then in the “life moves fast” category before I could finish any of these thoughts, DoNotPay made the “big” announcement today that it was not only abandoning its current plan to have its AI tell a pro se litigant exactly what to say in an upcoming court proceeding using Air Pods but also essentially shutting down all of the kinds of services that Tewson was publicly demonstrating don’t work and are not really even AI products as advertised.

To get the feel of that … here’s DoNotPay’s announcement:

Joshua Browder on Twitter: “We have some incredibly exciting announcements regarding GPT consumer rights products in the next two weeks. I have realized that non-consumer rights legal products (e.g defamation demand letters, divorce agreements and others), which have very little usage, are a distraction.” / Twitter

And here’s the follow up thread from Tewson:

Kathryn Tewson on Twitter: “The “exciting announcement” is that he’s getting rid of all the legal products on the site, including the very two products I was waiting to see my examples of!” / Twitter

All of which is to say (I think) that today’s takeaway ought to be that focusing on the ways that advances in AI are going to be able to be used by lawyers to make their jobs more efficient and push the costs for legal consumers downward is a much better approach (at least in today’s landscape) then fearing that “robot lawyers” are going to replace lawyers made of flesh and bones.

Leave a Reply

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

This site uses Akismet to reduce spam. Learn how your comment data is processed.