Courts across the United States are being buried under a growing tide of AI-generated filings, deepfake evidence, and hallucinated legal pleadings with no end in sight.
Since April 2023, Damien Charlotin, a senior research fellow at HEC Paris, has maintained a database tracking cases involving AI misuse in legal proceedings as they occur.
As of July 29, 2026, that database has recorded a total of 1,812 cases, with 909 of those occurring in 2026 alone, representing a dramatic acceleration in the problem.
Whether it involves deepfake evidence or fabricated citations, AI-generated falsehoods are consuming court time and draining resources that judicial systems can ill afford to lose.
The core problem facing courts is that the technology underpinning these fabrications is advancing far more rapidly than the detection tools and regulatory frameworks designed to counter it.
For every deepfake exposed and punished, countless others may slip through undetected, forcing courts to devote ever-increasing resources to separating reality from convincing fiction.
A Kentucky appeals court case decided in May 2026, Adams v. Anderson, No. 2025-CA-1309-MR, illustrates precisely how AI fabrication is now infiltrating even family law proceedings.
In that divorce case, an appellant seeking to modify a custody arrangement submitted supporting documentation including a photo of a prescription bottle and medical records purporting to show a child’s urgent care visit.
The appellee contended that he had contacted the urgent care facility directly and found no records supporting the visit, arguing the evidence was “fake, photoshopped, or created by artificial intelligence.”
During a subsequent contempt hearing ordered by the court, the appellant admitted she had lied and that the supporting documents submitted into evidence were fake or altered.
Courts do retain tools to address this misconduct, including sanctions, case dismissals, and in serious instances, incarceration, but these measures have so far failed to deter bad actors.
As one commentator noted, “I’m afraid that the message to stop using AI improperly will fall on deaf ears. After all, attorneys haven’t listened, so why should litigants?”
The concern is well founded given that monetary sanctions imposed on legal professionals for AI misuse have repeatedly failed to curb the behaviour across the profession.
As AI output grows exponentially more convincing, the burden placed on already stretched judicial systems will only deepen, raising urgent questions about how courts can effectively adapt.