Invented citations
-
When AI-prepared work goes wrong Cases and sanctions
An Orange County Appeals Court Sanctions a Supervising Lawyer Who Says He Reviewed a Petition's Arguments but Not Its Citations
In an order of August 2026, the appellate division of Orange County's superior court in California sanctions two lawyers, one of them the supervising attorney, who did not review the citations, over a petition whose table of authorities listed four cases that "appeared to have been fabricated" by generative AI.
-
When AI-prepared work goes wrong Cases and sanctions
California's Office of Tax Appeals Says an Appellant Cited Opinions of the Tribunal That Do Not Exist
In a precedential opinion issued on May 15, 2026, California's Office of Tax Appeals says an appellant in a sales tax case cited opinions of the tribunal that do not exist, and that parties and representatives appearing before it "are expected to know about the risk of AI hallucination."
-
Evidence and documentation Documenting AI-prepared work
Australia's Fair Work Commission Says People Who Use Generative AI to Help Prepare a Document for One of Its Cases Must Say So, Check It and State That They Did
A guidance note of the Fair Work Commission, published on August 24, 2026 and commencing on October 20, sets requirements for anyone who uses generative AI to help prepare a document to be lodged in one of its cases, and says the checking "must be done by a person."
-
When AI-prepared work goes wrong Cases and sanctions
The Solicitors Disciplinary Tribunal Struck Off a Lawyer Who Filed Misleading AI-Generated Citations and Repeated the Errors After They Were Pointed Out
In a judgment dated August 25, 2026, the Solicitors Disciplinary Tribunal found that a registered foreign lawyer had filed misleading quotations and citations "in consequence of" his use of generative AI, and had sent further erroneous ones after the errors were pointed out. It ordered him struck off, on this and a separate allegation. The decision is subject to appeal to the High Court, and the order remains in force meanwhile.
-
When AI-prepared work goes wrong Liability and insurance
The SRA Warns That Solicitors and Regulated Individuals Who Use AI Remain Accountable for Their Work, Regardless of How It Was Prepared
In a warning notice published on August 17, 2026, the Solicitors Regulation Authority said that solicitors and regulated individuals who use AI in delivering legal services "remain accountable for their work and outputs, regardless of how that work has been prepared," named two areas of particular concern, false or incorrect information in court and other documents, including citations, and client confidentiality, and said that those who do not have proper regard to the notice are at risk of disciplinary action.
-
Cannot Be Delegated to Software: The Texas Accountancy Board on AI and the CPA, August 2026
In its report of August 2026, the Texas State Board of Public Accountancy says a Texas CPA "remains responsible for all work product" and that "Professional responsibility cannot be delegated to software." The state's Public Accountancy Act and the board's Rules of Professional Conduct "do not specifically address generative AI," it says, but "existing professional obligations apply regardless of the technology used."
-
Regulators and standard setters Tax
Not a Substitute for Professional Judgment: The IRS Alert on AI in Tax Practice of June 2026
On June 24, 2026, the IRS's Office of Professional Responsibility issued introductory guidelines on AI in federal tax practice. Due diligence, it says, "requires verifying the accuracy of facts, citations, and calculations produced by AI."
-
Regulators and standard setters Tax
A Less Experienced Junior Colleague: Seven Professional Bodies on AI in UK Tax Work, January 2026
In January 2026 the seven professional bodies that jointly prepare Professional Conduct in Relation to Taxation, the principles and standards their members follow in tax work, said that AI output "should also be regarded as if it were prepared by a less experienced junior colleague and reviewed with appropriate scepticism," and that members remain ultimately responsible for their work "irrespective of the use of AI tools in its creation."
-
When AI-prepared work goes wrong Liability and insurance
Whether They Intend to Cover: A Lloyd's Market Association Report on AI and Errors and Omissions Insurance, September 2025
A report of a committee of the Lloyd's Market Association, published on September 18, 2025, says that a professional who relies on generative AI is "still expected to adhere to the professional standards" of their regulator, and that underwriters "should consider whether they intend to cover losses resulting from reliance on inaccurate GenAI outputs by professional services firms."
