Initial Thoughts on the Innovations in Technology Conference

Really this is about having too much AI content

The obligatory robot picture

Having both a pre and post-conference on AI was a bit much:

Because I’m me, I signed up for both the pre-conference on AI and the post-conference on AI, sandwiching a conference where I’d guess half of the sessions had something to do with AI. AI and GPT you say? Ay I’m Gettin Pretty Tired of, ok, I’ll stop.

The pre-conference (on AI):

I hate to say it, but the pre-conference on AI that LSC put on felt very much like a continuous sales pitch for CoCounsel, the Thomson Reuters AI product that they bought from Casetext at great expense.

In the midst of the multiple sales pitches for CoCounsel, a couple of things struck me:

• Why are we trying to automate interesting legal research and not the boring stuff that lawyers have to do? In fact, two out of three of the audience questions after the sales pitch were “how would I use I to automate [boring thing] that I have to do inside [matter management system].” I’d like to see a half-day on how to automate boring stuff with AI, but I am a very dumb person.

• The presentation on using OpenAI + Zapier + Airtable to automate expungement evaluations and paperwork was really good, and very practical. Granted right after that the TR salesperson gave a presentation on … how to automate expungement evaluations with CoCounsel.

• Do the majority of legal aid organizations really have a need for these erudite legal research or e-discovery tools that were originally built for AmLaw 100 firms? I don’t think the use cases are the same, i.e. most legal aids aren’t going to take complex federal litigation that requires e-discovery or high-value motion practice. Granted some will, like the few that do impact litigation and appeals, but that’s a teaspoon in a bucket. The vast majority of, say, legal aid eviction or domestic violence cases don’t exist in the world that most of the lawyers and technologists at Thomson Reuters exist in.

• This whole half day felt like cheerleading to an audience of very mixed opinions on even the concept of using AI for legal work, for various reasons. When your presentation is all about how awesome AI is, and the audience is asking questions like “well what about the devastating environmental impacts,” I’m not sure that your cheerleading is having the right effect. It comes off as arrogant.

• Ransom Wydner was the last presenter and did a walk-thru on how to create a custom GPT. This was a good session (Ransom is also a great presenter), and it should have been first or second on the agenda.

• I was probably not the intended audience for this half-day of AI fun, as evidenced by the above.

The third question from the audience in one of the sales pitch sessions basically boiled down to: “If TR really believes in the transformative power of AI for access to justice, why are their AI tools only available to lawyers?” The salesperson that TR sent to talk about their product didn’t really have an articulable answer for this. I’m of a mixed opinion here - I’m all for letting people use technology and against paternalism, but I’m also a believer that certain tools aren’t things that the general public should have access to. An example from healthcare: should the public have access to an MRI machine and get to self-diagnose? Granted the giant-ass magnet in an MRI could probably do some spectacular damage with a few misplaced paperclips (if movie and TV shows are to be believed), while legal research tools are somewhat more benign. But I don’t think TR has an obligation to make a tool designed for lawyers available for public consumption, especially considering that the public has access to OpenAI / Anthropic / Gemini.

The post-conference (on AI):

I had to miss a good bit of the post-conference and really regret it now, because this one was more of an academic conference (and I didn’t feel like I was being sold something). You can view the agenda here. Some takeaways:

• It’s really nice when panelists disagree with each other. This may sound weird, but the point of having a panel of people who know what they’re talking about is to get interesting and different opinions from knowledgeable people. There was one panel moderated by Jason Tashea where the panelists really got into disagreements (in a professional way) and everyone was able to articulate their viewpoint. This was a welcome contrast to the pre-conference where it felt very homogenous.

• If I heard correctly, Gillian Hadfield mentioned in passing that OpenAI had gotten letters from a bar association - if anyone has any info on this, I’d want to see it because it’s very interesting.

• Margaret Hagan raised the issue of what we want legal work to look like 5 years from now, assuming AI is by then capable of doing a lot of what lawyers do now. Are lawyers going to be “AI babysitters” and just reviewing the outputs for accuracy? Or are they going to be able to “practice at the top of their license,” whatever that may mean for them? I think this deserves more thought than what I’ll give it here, but I think of how pharmacists practice now versus 50-60 years ago - in retail pharmacy, the pharmacist is there to ensure accuracy of outputs (right number of the right pills in the right bottle). 50-60 years ago pharmacists did true compounding, where they needed to actually create the prescribed medicine from a list of ingredients, think about things like titration, absorption rates, etc. Some pharmacists still do this, but not the vast majority. Maybe we’ll see a future where we have “retail law” and “bespoke law.”

The Beagle+ folks were impressive:

The folks at the People’s Law School in British Columbia came and gave a presentation at the main ITC conference about Beagle, the AI bot they have rolled out. You can watch the Beagle+ session here. I was really impressed by their practical approach, especially their willingness to test and evaluate the bot’s performance. Also they have been running QA on the bot’s responses and documenting it - this will be really valuable for other organizations that want to do something similar. Granted most orgs won’t have the people power to do as extensive QA, but they can point to Beagle’s results as a type of level-setting.

Errata:

• I did moderate a session, which you can watch here. Fortunately I did very little talking, which IMO should be the goal of every moderator.

• I really like this conference (despite my earlier grousing about it) mainly because I get to see people in person. I really miss working in an office most days.