The Volokh Conspiracy
Mostly law professors | Sometimes contrarian | Often libertarian | Always independent
"Lest We All Drown in a Sea of Slop"

From Kelleher v. Town of Brookfield, decided earlier this month by Judge Brian Murphy (D. Mass.):
Plaintiff's Complaint is more than 100 pages long, often repeating itself or presenting information piecemeal and out of order, to some sort of titanic effect. While the Court appreciates that there is a long history between these parties and that Plaintiff is a pro se litigant and so entitled to a fair amount of leeway, "[e]ven pro se litigants are bound by the Federal Rules of Civil Procedure," including Rule 8's requirement that pleadings be "short" and "plain," with allegations that are "simple, concise, and direct."
Of course, the Court is well aware that attorneys, too, have long and often violated this rule (sometimes at great profit). See, e.g., Trump v. New York Times Co. (M.D. Fla. 2025) (excoriating experienced counsel for submitting a complaint that was "decidedly improper and impermissible"). However, particularly as artificial intelligence makes the production of language cheaper and faster—undoubtedly a boon to those that have historically been unable to afford garrulous counsel—it will become increasingly incumbent upon courts to insist that parties respect our limited bandwidths, lest we all drown in a sea of slop.
The Court would not make an example out of this case. However, moving forward, the Court would ask that Plaintiff (or any other prospective filer) do his best to write plainly and precisely, trusting that the Court reads closely and is unmoved by too much rhetoric.
Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of Reason.com or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Comments may only be edited within 5 minutes of posting. Report abuses.
Please to post comments
Some judge out there must have good enough AI skills to produce a video of summary execution of a lawyer in his courtroom for filing an AI-assisted brief. Maybe that would be unethical?
I know of students who were committed to the psych ward for stuff like that, I’ve attended trainings where we’ve been warned to watch for stuff like that.
To the extent that anyone is watching the mental health of judges, this would be the fire siren in the night. Like I said, to the extent…….
As an aside, anyone remember what happened when Sarah Palin put someone’s picture on a pistol target? This is a wee bit more extreme…..
The dead lawyer doesn't have to be real. The point is to let the profession know judges are mad as hell and they're not going to take this anymore.
I argue that this is an example of why the law school prerequisite to the bar should be abolished, and for good reason.
First, in discussing a bit of litigation that I was incidentally involved in with an attorney friend, I asked why opposing council — well regarded and quite expensive — took over three pages to say something I could’ve said in a paragraph, that should’ve been said in a paragraph
She pointed out that such was an acquired skill that one learned in law school.
And second, perhaps more importantly, it is the high price of council that forces clients to go pro se. and the high price is the consequence of one factor, the high opportunity cost of law school.
I noticed on the mass municipal association website that Brookfield is looking for an interim town treasure, and an interim town administrator — in addition to people to hold those jobs full-time. I’ve seen other jobs posted in Brookfield recently.
The only reason to look for an interim is that the incumbent either left without notice or was abruptly fired, sometimes a distinction without a difference.
Looks to me that it’s a nice little cauldron in that town….
Who knew that equipping impecunious pro se litigants would provide an upside to offset the flood-the-zone advantages rich clients buy when they hire big law firms. Yikes! The world turned upside down.
AI does mistakes but the mistakes are becoming fewer and fewer but most importantly AI allows skilled users to pose many what-if scenarios which can assist in presenting an argument that will do well in court.
AI has people scared they have to latch onto the few outlier spectacular failure of its use. The point is before the decade is out AI will be better than anyone can imagine and will exceed limits many fear. It is just an application of power and those limits are being expanded upon continuously
This is an idea I have been espousing for some time, to no avail. One point often overlooked is that what I will term "free AI" combined with poorly formed prompts provides just what the prompter asked for, but not what they wanted. Even low level 'pay for tokens' LLMs are limited.
Not trying to introduce too much nerd talk but I just installed Ollama which is required to run some flavor of Llama. The question is which flavor of Llama, I chose Llama3-Abliterated. Next question is how big, I chose Llama3-Abliterated72b realizing I will only get a couple of words a second but a better answer while using the 32b provides a massive upgrade in speed but at the cost of a top tier analysis.
But once I have this top tier LLM running locally on my dedicated AI box there is still the issue of the prompt. I normally use a text editor to create a rough prompt. Next I add a disclaimer at the front along the lines of 'this is a rough prompt giving you some idea of what I am looking for, please edit/rewrite it to improve the results of what I want' with the same disclaimer at the end, adding that I will likely revise the prompt after I see the result. As a rule the AI rewrite of my prompt is 5-10 times longer than my original prompt.
This is not the first I have posted this as well as my 1L story. First class, first instructor said "I expect anything you turn will have been edited at least ten times before you turn it in". Point is if the same effort is used to create the prompt as to create assignments/brief AI works; if not it is the old garbage in, garbage out.
AI is like any other tool -- it has the potential of great usefulness, but it also has the potential of causing great injury if you don't use it properly. Of course AI can help a skilled user. But most people aren't that skilled, and using it badly can lead to a lot more harm than not using it at all. I don't advocate "no AI under any circumstances," but acting like it will somehow solve all your problems on its own doesn't do anything to address the problems.
The mistakes are not becoming fewer and fewer.
And lawyers would be 'skilled users' by any useful definition of the term and they keep turning in AI slop.
You can say that 'before the end of the decade' - but that's 4 years away and we live in *today*, not tomorrow.
And today AI is garbage.