From Traditional Drafting to AI-Assisted Drafting: Why Attorneys Need a New Standard of Review for Generative-AI Work Product
Legal practitioners are increasingly using AI tools to improve efficiency and quality of legal services. This change develops from the growth of artificial intelligence, especially generative AI, (GAI or Gen AI), from a technological trend into a practical tool in legal space. Lawyers use Gen AI for research, draft pleadings, motions, contracts, discovery and other litigation support documents, including client communications and first draft analysis. The growing popularity of AI tools stems from a simple reality. Legal drafting is time consuming, detail oriented, and often repetitive. This is where Gen AI gives clear value. It can help produce faster outcomes, organize clear arguments, identify drafting issues at an early stage, and support attorney judgment, legal strategy, and review.
Is Gen AI just a faster writing assistant? No. Specifically, in legal drafting, it could create unique professional risks. The outcome may be fluent, plausible, and at the same time flawed too. As a result, AI-assisted legal drafting might be a challenge for an attorney’s professional duties under the American Bar Association Model Rules. Attorneys, therefore, require a new standard of review or AI verification protocol for their internal processes to ensure compliance without compromising their professional and ethical duties towards clients and society.
ABA Model Rule of Professional Conduct 1.1 provides: “A lawyer shall provide competent representation to a client” and that “[c]ompetent representation requires the legal knowledge, skill, thoroughness and preparation reasonably necessary for the representation.” Likewise, the Federal Rule of Civil Procedure 11 requires attorneys and unrepresented parties to verify and sign court documents and certify that the document is filed for a proper purpose, contains legally sustainable arguments, and have evidentiary support for factual contentions. A violation of the rule can result in monetary or disciplinary sanctions. This is when the widespread use of Gen AI becomes significant.
Recent data tracking U.S. court filings involving false legal information—including hallucinated citations, incorrect reasoning, unverified holdings, and fabricated quotations— shows the inherent risk of using Gen AI in litigation.1 The problem is not just related to AI creating errors, but that those errors may appear professional, genuine, authoritative, and citation-supported unless an attorney independently verify the results before filing in court.
In a recent case, LNU v. Blanche,2 the Ninth Circuit, sanctioned attorneys after appellate briefs contained nonexistent cases, wrong quotations, and serious misstatements of real authority. The court made clear that it was not punishing the mere use of Gen AI. The attorneys violated their conduct when they signed and filed the briefs without adequately verifying the authorities in it. According to the Court, an attorney’s signature certifies personal responsibility for the filing, regardless of whether the draft came from a subordinate, AI tool, or the attorney’s own work. The court, here, emphasized candor i.e., once attorneys discover hallucinated citations or such errors, they must promptly disclose the error and its source without reframing it as a typographical mistake. The attorneys received monetary sanctions, a fine of $2,500 each, a six-month suspension from Ninth Circuit practice, and future AI-disclosure and citation-verification requirements.
Similarly, In re Nwaubani, No. 25-9517, the Fourth Circuit, publicly admonished an attorney for citing three nonexistent cases in his appeal brief in an employment discrimination case. The court did not decide whether Gen AI produced false citations. Rather, it made a broader reasoning to find that the submission of the brief with nonexistent authority violated an attorney’s professional obligations regardless of the source of the error. Citing Fed. R. Civ. P. 11, the court highlighted that attorneys are gatekeepers “to ensure the accuracy of their filings.” The court further emphasized that “[a]ttorneys have ethical obligations as to the materials and arguments they advance in federal court.” Even though “making a mistaken citation in a brief does not automatically rise to the level of a Rule 8.4(d) violation,”3 citing “three nonexistent cases, each with materially different citations than those of the cases” is a violation.4 The court noted that “[a] reasonable attorney, in preparing or reviewing his brief, should have discovered the errors.”5
These decisions emphasize a practical rule for AI-assisted drafting. Gen AI may assist attorneys in several tasks, but attorneys have an ethical obligation to independently verify every citation, quotation, and legal proposition before filing it with the court. An update on Gen AI sanctions in 2026 shows that use of such tools without properly verifying the AI generated output is becoming a common practice among lawyers. A top tier law firm, Norton Rose Fulbright’s 2026 update, AI in Litigation: Update on Gen AI Sanctions in 2026, reflects the intensity of the problem.6 The article reports that, “a researcher who maintains a database tracking Gen AI-related court orders in the United States has documented over 1,148 cases of hallucinations by lawyers.”7
Norton Rose Fulbright’s analysis of such recent decisions involving AI hallucinations showed a pattern of “how an attorney responds when confronted with Gen AI hallucinations in a filing can be just as consequential as the underlying error itself.”8 The article notes that US courts are specifically crediting “candor and cooperation” as mitigating factors and are imposing heavy penalties on attorneys who try to “evade, mislead or challenge the legitimacy of the court’s inquiry.”9 At the same time, the article cautions that even “an honest response does not guarantee immunity from serious professional consequences” because “the gravity of the underlying misconduct remains independently relevant.”10
Although there are no uniform nation-wide rules in the US governing the use of AI in courts, several statewide judicial directives, rules, regulations and standing orders from individual judges and state bar ethics provide guidelines on the same. The ABA also issued its Formal Opinion 512 to address the ethical responsibilities attorneys must consider while using AI in their legal practice. The opinion does not treat AI as prohibited or risk-free. Rather, the opinion describes it as a useful tool that requires lawyer supervision. ABA says that attorneys may use Gen AI for legal tasks including drafting, but they are responsible for “their duties to provide competent legal representation, to protect client information, to communicate with clients, to supervise their employees and agents, to advance only meritorious claims and contentions, to ensure candor toward the tribunal, and to charge reasonable fees.” The message is clear. Gen AI can improve speed and efficiency, but it cannot replace an attorney’s independent judgment, review, verification, and professional responsibility.
Here, to comply with ABA Rule 1.1., attorneys must understand the benefits and risks of using Gen AI competently, rather than blindly relying on the outcome. ABA Rule 1.6 requires attorneys to protect confidentiality of client information when inputting prompts, documents, or facts into AI tools. Rules 3.1 and 3.3 require attorneys to not bring or defend a frivolous claim and maintain candor towards tribunals. They must ensure to bring only legally sustainable arguments with proper citations, quotations, and accurate factual representations. Rules 5.1 and 5.3 impose ethical obligations on law firm partners and supervising lawyers to adopt reasonable measures to ensure that lawyers and non-lawyer staff comply with the rules of professional conduct. Finally, Rule 1.5 requires attorney billing practices to reflect efficiency gained from using AI, rather than charging clients for time not actually or reasonably spent.
In addition, states and local courts are independently prescribing rules on using AI in legal practice. For instance, the Kansas Rules of Professional Conduct, following the Model Rules, prescribe similar duties of competence, including technological competence. See Kan. Rules Pro. Conduct r. 1.1 cmt. 8 (2025).11 Johnson County, Kansas Local Civil Rule 3.7 requires the parties and counsels to expressly certify whether Gen AI was used in preparing any court document. The rule emphasizes that attorneys are responsible for the accuracy, quality, and legal sufficiency of all pleadings, when using AI assisted tools in drafting. The Rules also identify the risk associated with Gen AI tools and require attorneys to independently verify and review every citation or law on the record, quotations, paraphrases, assertions or legal analysis contained in the documents filed with the court. When Gen AI is used, the certification must identify the AI tool and confirm compliance with K.S.A. 60-211 and applicable ethical rules. An incorrect or missing certification may result in rejection of the document or sanctions.
However, disclosure rules might not be a complete safeguard for Gen AI assisted legal drafting. Such disclosures may alert the court to the potential risks in the filing, but they do not, by themselves ensure that the final work product is accurate, non-frivolous, confidential, and consistent with the attorney’s duties of candor and competence.
The question is not whether an attorney used AI tool for legal drafting, but whether the attorney used reasonable care to verify the work before filing it with the court on behalf of his client. Tasks such as formatting and stylistic edits may require only ordinary review. However, legal research, factual statements, contracts, and legal arguments require more thorough review. Law firms should adopt practical protocols or review procedures to verify sources, quotation and record citations, while using prompt playbooks, best practices, and heightened review for high-stakes work. With appropriate standards and protocols, AI tools improve efficiency and consistency while reducing the risk of malpractice claims, reputational harm, and sanctions. The advantage will not belong simply to those attorneys who use AI tools, but to those who know how to use them responsibly and well.
By Vinitha Prasannan, Senior Manager – Legal
For more practical litigation drafting tips, pleading and legal research insights, follow LawCompany.