When lawyers started using ChatGPT to write their papers, resulting in hallucinated cites and non-existent quotations, not to mention arguments premised on the inability to grasp basic legal principles like jurisdiction, they got caught, bench-slapped and often sanctioned. I thought they would learn. I was wrong.

I called these lawyers lazy and greedy, unwilling to put in the time needed to craft winning papers, but all too happy to charge clients as if they did for work they didn’t perform. Hell, they couldn’t even be bothered to check their papers written by AI, which could have saved them the humiliation of a judge calling them unpleasant names. And yet, lawyers continued to use AI, including biglaw, charging shamefully high fees for baby lawyers for work that was farmed out to the computer.

Learn from mistakes? Oh no. Then judges started using AI to write their decisions. Perhaps it was their law clerks who did the dirty, and judges were too busy to check the work and signed off on it, even though it was AI slop with the same hallucinations for which they would hand a lazy lawyer their butt. Surely, this teach the legal profession the error of their artificial ways, right? Oh, sweet summer child, hardly. There was still other demands of the system that had yet to aspire to easy-peasy failure. Enter the court reporters.

Although “[t]rial records are rarely if ever perfect,” Ben-Yisrayl v. State, 753 N.E.2d 649, 662 (Ind. 2001), the Transcript in this case is far from the best.

The Transcript contains various types of errors. There are numerous typos that change the meaning of the testimony, question, or objection. See, e.g., Tr. Vol. II at 137:18, 144:10, 147:10; Tr. Vol. III at 6:13. In some instances, witnesses’ and trial attorneys’ names are reported incorrectly. Tr. Vol. II at 220:5; Tr. Vol. III at 142:15–20, 143:15, 162:4–5.

At one point in the Transcript, a motion, presumably made by the State, is attributed to the trial court. Tr. Vol. II at 107–08. At another point, an objection, presumably made by Williams, is attributed to the Bailiff. Tr. Vol. II at 177:15. At yet another point, the State’s closing argument is attributed to the trial court. Tr. Vol. III at 228:1.

The Indiana Court of Appeals in Wiilliams v. State didn’t find the errors in the transcript so bad as to impede the court’s review of the underlying case, which involved the sale of drugs and resulted in a 40-year prison sentence, but it was bad enough to be worthy of mention in the footnotes.

Based upon the types of errors reviewed, it appears that generative artificial intelligence may have assisted with the preparation of this transcript. While AI can improve efficiency and be a productive tool for many professionals, it is incumbent upon those using such systems to proofread and ensure the accuracy of the generated product.

Rather than condemn the use of AI, the court extolled its virtues with a cautionary warning. As with lawyers who use AI to do their work, it cautions court reporters and lawyers to “proofread and ensure the accuracy of the generated product.” In the ordinary course, lawyers are always supposed to proof and correct any errors in transcripts. Many don’t bother. Many have no memory of what exactly was said or by whom, and are incapable of correcting the transcript. And when a new lawyer does the appeal, she would have no ability to know what in the transcript was accurate and what was not. Are they supposed to rely on the trial lawyer, who was just fired and replaced?

For the most part, court reporters are pretty good at their jobs. At least they used to be. Sure, there was errors, but they tended to be very minor and easily corrected if they were material to the case. But court reporting was a lot of work, and typing up a transcript was both an expensive and time-consuming project, necessary for appeal though it may be.

In the future, creating a transcript will be fast and easy because the work won’t be performed by busy fingers, but by AI chatbots. Why should the fact that the next 40 years of your client’s life depends on the accuracy of the transcript concern you? Isn’t “efficiency” and a “productive tool” more important than getting it right?

As was noted years ago as technology sought to wiggle its way into the law, just because it makes things easier does not mean it make things better.

It’s not that lawyers are anti-technology, it’s that they are anti-bullshit.

— Keith Lee

Remaining anti-bullshit seems prudent, even if generative artificial intelligence persists on pushing lawyers down the path of incompetency.