The Alaska Files: Senate Judiciary 02/15/23

Well, this is now officially a series. Here I intend to give some thoughts on the Alaska Senate Judiciary Committee meeting of February 15, 2013, which can be viewed here. Once again, this is not an analysis per se, just some of what I was thinking while I listened to the meeting.

The focus of this meeting was on a presentation given by Deputy Attorney General John Skidmore, which seemed primarily concerned with sort of the internal health of the Department of Law: how the department managed setbacks caused by COVID, struggles with employee retention and new applicant quality, just how the department is currently functioning, and how it hopes to function in the future.

But two senators, Kaufman and Tobin, each made speculative statements that sort of pull in opposite directions while trying to get at the same goal, and it is those statements I want to explore.

Senator Kaufman’s statement came first: the gist seemed to be that he wondered if utilizing AI might alleviate some of the department’s staffing issues. Mr. Skidmore acknowledged that he had no expertise in the area of AI and hadn’t even looked into that as a possibility. Technically I don’t have much experience with AI, either. But I have played with some advanced AI, and I do have opinions.

Current AI has one massive flaw: it is not designed to create true things, it is designed to create things that appear true. It isn’t a huge stretch to say that current advanced AI is designed to lie convincingly. It utilizes massive databases of inputs to determine what an answer to a particular query should look like, and then it fills in the blanks.

For a lot of applications this isn’t a huge deal. But for the purposes of law, and especially for the purposes of a government lawyer, this is critical. If one of the pieces of data that the engine draws upon cites a law from the wrong state, a human might not catch that difference, and that can have a massive impact on a court case. Entire documents may be thrown out.

I can imagine a scenario where a lawyer is up against a deadline and can’t focus, and so uses AI to generate a template that the human then uses as a basis to manually create their own document from scratch. But I think it’s important that every word of the final document be written by a human, be created by a human brain. And when you strip down AI processes enough so that no unintentional shenanigans enters a legal brief, essentially what you’ve done is created a Word template. So you might as well just have a database of Word templates that you reference when needed.

In short: I think we are still very far away from a time when AI can be reliably used anywhere in the legal profession.

But while Senator Kaufman’s statement seemed to encourage automation, Senator Tobin’s statement did the opposite: she speculated on a future where the bar exam was no longer a standard requirement for practicing law.

This wasn’t within the purview of the meeting, I don’t think, but the first question that comes to my mind is: what replaces it? That isn’t necessarily a critical question, it’s just a practical one: the replacement may be “if you’ve earned a JD, then that’s sufficient to practice law.” The replacement may be some stricter licensing requirement. Or some proof of qualification that is more creative than an exam. Maybe (and I am not a lawyer), rote knowledge of facts isn’t the best measure of the quality of a lawyer in 2023, even if it may have been in the past.

While this statement, I think, is more feasible on a strictly practical level than the introduction of AI, and in theory it could be implemented within the next decade if everyone decided tomorrow that it should be done, I also think the legal profession is Very Set In Its Ways, and such a change will grate on those who move the levers of power that may implement it. It may take this generation of lawyers maturing up to the point where they are in those positions before this happens.

But both of these are new ideas. We have to process through them fully before we decide whether or not they are good ideas. Tomorrow there may be a new development in AI technology that completely invalidates everything I’ve said here. Or the movement to abolish the bar exam may swell in numbers and its sunset will seem inevitable by the end of the year.

But, while interesting to consider, I don’t see either of those ideas as viable solutions in the short- or medium-term.