Indicia of Misuse of A.I. – Courts Need Not “Fish a Gold Coin From a Bucket of Mud”

No Violation of A.I. Protective Order
September 11, 2026

Defendant Lucas Sirois sought to overturn his conviction based on alleged discovery of new evidence.  He filed an 80 page motion, with a proposed 101-page declaration, “which works in tandem” with the lengthy motion. Overall, the motion “approaches 200 pages in length.”

The court wrote: “The proposed filings are clearly not the work of an organic mind with linguistic talents developed in the crucible of human experience. It is obvious that the entity that drafted them has an artificial intellect that is unconcerned with, for instance, getting to the point. To comment on the filings at length would only serve to reward them. Still, I offer some particularly glaring artificial ticks that are a complete waste of the reader’s time. Throughout the motion, the author has taken special care to tell the Court what it is NOT arguing or suggesting.”  U.S.A. v. Sirois, 2026 WL 2671655 (D. Me. Sep.11, 2026)(emphasis added).

The court provided 10 examples and wrote: “Many, many more such empty statements follow. There are equally unproductive positive descriptions of what the Defendant purports to do with the motion, as though it were necessary to narrate the action rather than simply present the argument.”

The court stated: “Although counsel is free to employ generative artificial intelligence (‘AI’) to assist with the production of motions and other appropriate papers, the Court expects that counsel will distill the resulting artifacts to their essence in order to present filings that are accurate and concise and in keeping with ordinary practice.”  It added:

If every exhibit must be scanned and uploaded for critical commentary by AI, so be it, but counsel may not supplant the record with an artificial facsimile and force the consumption of AI circumlocutions without end. If these sorts of filings pass muster here then they must pass muster in every other case. That development would mark the end of human-centered jurisprudence because it will prove necessary for judges and their staff to outsource their own duties to an AI tool. In short, the Defendant and his counsel must draw upon whatever lessons and insights these tools may impart and attempt to translate them into a set of much more compact and cogent contentions that are more readily digestible by a human decision maker. [emphasis added].

In another recent decision, the court wrote that it “need not try to fish a gold coin from a bucket of mud.”  Some Unacknowledged A.I. Comes Within the “I’ll know it when I see it” Principle (Jul. 1, 2026).  I guess that is another way of saying that “Judges are not like pigs, hunting for truffles buried in the record.” (Mar. 21, 2025).  For another example, see Nonsensical Spellings and Fabricated Authority Signal Improper Use of Artificial Intelligence (Mar. 25, 2026).

Share