It Will Be Your Name and License on the Line, Not Chatgpt’s

From Judge Kai Scott (E.D. Pa.) today in Bunce v. Visual Tech. Innovations, Inc.:

… Plaintiff … sought reasonable travel costs for a cancelled deposition. In response, Mr. Rajan filed an omnibus motion seeking sanctions and objecting to the Plaintiff’s travel costs. Plaintiff responded accordingly, highlighting that although Mr. Rajan was sanctioned previously “for his use of made-up cases and other authorities that did not support his propositions[… ] Rajan has once again used a made-up case and cited to other authorities that don’t support his propositions[.]” In his Reply, Mr. Rajan only briefly addressed this issue in a footnote.

On March 11, 2026, this Court Ordered Mr. Rajan to pay Plaintiff all reasonable travel costs, denying his omnibus motion and ordering him to show cause as to why his citations in his omnibus motion did not violate Rule 11(b) or this Court’s Standing Order Re: Artificial Intelligence (“AI”). Mr. Rajan filed a response to the Court’s Order to Show Cause in which he again responded to the claims of improperly using AI….

Contrary to his assertions, it was patently unreasonable for Mr. Rajan to file his omnibus motion with erroneous citations, at least one of which was made up. In his Response to the Court’s Order to Show Cause, Mr. Rajan asserts that “[t]he question is whether a reasonable attorney, under the particular circumstances, would have acted similarly.” He answers his own question in the affirmative, stating that a reasonable attorney would have filed the motion with the erroneous citations because the opposing party refused to confer with him on the matter of travel fees.

This is incorrect, both as a characterization of the governing law and what a reasonable attorney would do. As discussed above, the proper question is whether Mr. Rajan should have known that the citations in his filing were incorrect. He should have. As Rule 11 discusses, every attorney and pro se litigant must verify the veracity of their filings before submitting them to this Court. Full stop. It was not enough that Mr. Rajan felt like he had no other choice because opposing counsel refused to meet with him to discuss a matter that the Court already resolved.

Nor was it enough that Mr. Rajan believed that his time to object was set to expire soon so he had to rush to submit his filing. He simply could have requested an extension of time, which the Court would have readily granted. Instead, he “haphazard[ly]” submitted a document generated by AI without verifying its accuracy and then, when challenged by opposing counsel, deflected by trying to blame opposing counsel, doubled down on his conduct, and argued that the Court’s Standing Order did not clearly specify the need for “human verification only.” This approach was ill advised.

Mr. Rajan was unable to identify any valid reason for why he had not verified his erroneous citations. He stated that the erroneous citations were mistakes due in part to the fact that he does not normally practice law. This explanation was unpersuasive. Any first-semester, first-year law student would know that a fundamental rule of lawyering-during litigation or otherwise-is to ensure that any authority to which a lawyer cites does indeed support the proposition for which he cites it. It therefore makes no sense that Mr. Rajan—a lawyer for nearly 40 years—failed to do so, even if he primarily focuses on “business stuff’ and relies on local counsel to manage court filings.

What’s more, in his Response and at the hearing on the Order to Show Cause, Mr. Rajan pleaded with the Court not to sanction him more harshly than “a modest deterrent penalty of $950” due to the more than $73,000 in sanctions he has paid to date, the travel expenses he has already been ordered to pay, and the “unreimbursed time spent preparing the Omnibus Motion.” Thus, he continues, those fees and the reputational harm he has already suffered necessitate a lower sanction for his second time violating Rule 11. This Court disagrees.

First, this situation is of Mr. Rajan’s own making. This Court may be one forum in which he is held accountable for his improper conduct, but the Court is not responsible for the reputational harm and myriad fees he brought on himself. Second, and perhaps more importantly, the other fees were for conduct separate from the issues before the Court now. Mr. Rajan cannot expect not to be sanctioned for additional improper conduct because he engaged in other improper conduct previously.

If anything, the series of improper conduct demonstrates Mr. Rajan’s inability or unwillingness to rectify his behavior, a sign that additional deterrence is needed. See Hearing Tr. (“I’m not sure what I’m supposed to say to that, Your Honor. All I ask is that whatever the court thinks is reasonable and fair to stop me from this nonsense.”) For his previous Rule 11 violation, this Court imposed a penalty of $2,500. The Court remains unconvinced that a significantly reduced penalty for a second Rule 11 violation for the same conduct would adequately remedy the repeated behavior. Thus, the Court imposes a penalty of $5,000.

The Court remains appalled by Mr. Rajan’s improper conduct. Rule 11 makes clear that any attorney or pro se litigant submitting any paper to any federal court has an obligation to certify the veracity of the representations made therein. Furthermore, Pennsylvania’s Rules of Professional Conduct also impose a duty of candor to the tribunal that mirror those sentiments. Those duties do not disappear solely because an attorney chooses to outsource his labor to AI.

Indeed, it becomes even more important for lawyers to check their work when using technology known to be inaccurate at best or hallucinate and make up cases at worst. Should Mr. Rajan engage in this improper conduct before this Court for a third time, this Court will not hesitate to sanction him once more and refer him to the Pennsylvania Disciplinary Board for further proceedings.

The Court does not seek to discourage lawyers from embracing new technology in their practices. But like with any tool, lawyers must exercise pristine judgment when using AI while administering their duties before this or any court. The sanctions discussed herein should serve as an admonition to all lawyers that the law is fundamentally a human enterprise under which lawyers serve as vital stewards of society. Do not get so comfortable that you forget your duties to clients, courts, or society writ large, because if you do, it will be your name and license on the line, not ChatGPT’s….

Note that Rajan stated that he checked the citations—but using “an AI-assisted legal research tool.”