“[U]sing AI encourages personal detachment from the attorney’s work product and relationship with the Court.”

In Davis v. Marion Co. Superior Court Juvenile Detention Ctr., 2026 WL 1408448 (S.D. Ind. May 20, 2026), the court took a novel view of artificial intelligence.

The all-too-common issue presented in Davis was whether the court should “personally sanction Plaintiff’s counsel, …, for including two non-existent legal citations in a brief responding to a motion to compel. The legal citations were hallucinations generated by an artificial intelligence (‘AI’) research tool.”

The court rejected the assertion that the Indiana Rules of Professional Conduct had been violated: “Those issues have been referred to the Indiana Disciplinary Commission by the Chief Judge.” Therefore, the court deferred to the state disciplinary process.

Instead, it relied on Fed.R.Civ.P. 11.  That is not unique.

However, the Davis court added: “A heavy caseload and compressed schedule do not excuse [the attorney’s] obligation to implement a quality control process at his firm that would prevent fake citations. Most attorneys suffer from too little time. This is not the problem. The problem is not being ready for the inevitable time crunch.”

The court saw a very different issue:

Unlike mere sloppiness, reliance on fictitious citations highlights a concerning symptom of the legal profession’s growing adoption of AI: a disembodied practice. In other words, using AI encourages personal detachment from the attorney’s work product and relationship with the Court.

The legal profession is personal. Every lawyer is personally licensed, personally appears in court, and personally signs pleadings on behalf of their client. We also assume that lawyers have personally read every case they reference and have incorporated that case’s holding into their corpus of expertise. But when the practice of law becomes increasingly mediated by ever-advancing technology—which, frankly, most of us do not understand— something essential to the profession is lost.

Bar and court admission requirements also reflect the personal nature of the legal profession. “Only members of the court’s bar may represent parties before the court,” our local rules say. Local Rule 83-5(a)(1). There’s a vetting process. Coupled with the general prohibition on ex parte communications, the admissions process creates a semi-sacred space between the bar and the bench. The public must trust this space and know that the facts and arguments considered by the Court truly belong to the parties. When judges read legal briefs, they enter that semi-sacred conversation with confidence that the arguments they are considering are arguments fairly and honestly presented by licensed counsel or pro se litigants—people who are members of the Court’s community and have a personal interest in the conduct of the litigation and the administration of justice in their affairs.

Hallucinated citations rudely interrupt this conversation. It’s jarring and disorienting. Decorum and trust are breached. At this point, the Court cannot immediately discern if it is receiving argument from the person who signed the brief or from an artificial intelligence the Court has never met and has not licensed to appear before it—a voice that does not belong to a person at all.

In sum, fake citations represent more than sloppy lawyering and cost more than the Court’s time. They reflect a tendency toward professional detachment. This in turn jeopardizes the conventions of our profession, as well as the public’s trust that the arguments and rationale considered by the Court are the facts and reasons of the parties, not artificial interlopers. Monetary sanctions are appropriate to deter similar detachment in others and encourage a renewed commitment to the duties of candor and diligence in an age of advancing technology.

Id. at *4-5.

The Davis court added:

One step toward sound and workable practices for generative AI is the full disclosure of its use. All parties to a conversation, even artificial ones, should be properly introduced. Civility and professionalism warrant nothing less. Working in the light of transparency encourages both the awareness of AI’s novelty and personal accountability for its use. Accordingly, this Court mandates that all counsel and pro se litigants announce their use of generative AI in any legal filing.

Id. at n.1. Whether you agree or not, and I’m not sure that I do, the Davis court’s analysis is interesting.

For a different view, see the Hon. Scott U. Schlegel, 🚀 A Call for Education Over Regulation: An Open Letter — Judge Scott Schlegel (judgeschlegel.com).  “In my humble opinion, an order specifically prohibiting the use of generative AI or requiring a disclosure of its use is unnecessary, duplicative, and may lead to unintended consequences…. The legal profession is already guided by stringent ethical standards and professional responsibilities….  A focus on education and adaptability within these existing frameworks would be a more effective way forward.”

See generally:

• A Court’s Standing Order on A.I.;
• Should Courts Use Standing Orders or Local Rules to Address A.I.?

In the evidentiary context, please see Suggested A.I. Rule – Suggested Amendment to Maryland’s Computer-Generated Evidence Rule.