Artificial Intelligence, Real Liability: Could Not Using AI Be Legal Malpractice?

MBHB Summer Associate Aurora Lavin-Peter co-authored this article.

Ever since artificial intelligence (AI) tools like ChatGPT became publicly available, much ink has been spilled about their ethical use. Articles highlighting examples of lawyers misusing AI tools are passed around like legal ghost stories,[1] warning lawyers of the risks of careless reliance on AI. However, it is myopic to focus solely on these cases to decide whether to adopt AI tools. AI tools offer undeniable efficiencies in everyday lawyering. From research to analysis to drafting, AI can make many aspects of legal practice more efficient and robust. Given these benefits, this article challenges traditional legal discourse on the adoption of AI tools by asking two questions: First, could failing to use AI tools constitute malpractice? Second, how can existing guidelines and standards for AI use be made clearer?

I. Could failing to use AI tools constitute malpractice?

Although AI is a recent development in legal practice, the legal profession has confronted similar questions about technological disruptions. The T.J. Hooper[2] illustrates how the law approaches conflicts between industry practice and technological change. The T.J. Hooper involved tugboats towing barges along the Atlantic coast.[3] The tugboats encountered a severe storm, causing the barges to break loose and sink.[4] Although radio receivers capable of providing weather forecasts were available, the tugboats were not equipped with functional radio receivers because radios were not a standard practice in the industry.[5] Judge Learned Hand held the tugboat owners liable, famously declaring that “a whole calling may have unduly lagged in the adoption of new and available devices,” and that “reasonable prudence” is to be measured not by industry custom, but by the courts.[6] Much like the tugboat owners held liable for sailing without functional radios, attorneys who reject AI tools may one day soon face negligence claims as the technology becomes the new industry standard, especially paired with human oversight to mitigate risk.

It is not just century-old case law that suggests AI non-use could constitute malpractice; the Model Rules of Professional Conduct (“Model Rules”) and 37 C.F.R. provisions reflect a similar sentiment. Under 37 C.F.R. § 11.101, practitioners before the United States Patent and Trademark Office (USPTO) must maintain competence. Similarly, Rule 1.1 of the Model Rules demands that attorneys maintain competence.[7] Moreover, Comment 8 to Rule 1.1 requires lawyers to “keep abreast of . . . the benefits and risks associated with relevant technology.”[8] Thus, in states adopting this comment to the Model Rules,[9] failure to keep pace with technology could constitute malpractice.

Courts are unsympathetic to lawyers who profess ignorance of technology. For example, in James v. National Financial, LLC,[10] the Delaware Court of Chancery was unmoved by an attorney’s claims that “I am not computer literate . . . I need a secretary to help me turn on the computer,”[11] stating, “[p]rofessed technological incompetence is not an excuse for discovery misconduct”[12] and “deliberate ignorance of technology is inexcusable.”[13]

A second example of when an attorney’s non-use of AI could constitute malpractice is based on Model Rule 1.2 and its USPTO counterpart (37 C.F.R. § 11.102). Under Rule 1.2, attorneys must abide by a client’s lawful objectives and consult with the client regarding how those objectives are to be pursued.[14] Increasingly, clients are establishing outside counsel guidelines that direct outside counsel to use approved AI tools for specified tasks to reduce costs and improve efficiency.[15] Under Rule 1.2, a lawyer who categorically refuses to implement such AI tools, especially when contractually required by client engagement terms, would be at risk of violating their ethical duties.

A third example of non-use of AI potentially resulting in a claim of malpractice arises from a lawyer’s duty of diligence under Model Rule 1.3 and 37 C.F.R. § 11.103. Under Rule 1.3, diligence is measured not merely by effort, but also by “promptness” in pursuing a client’s interests.[16] As AI systems become demonstrably more efficient and accurate at reviewing large document collections, exclusive reliance on manual review will become difficult to justify.[17] Consider the scenario where a patent attorney manually reviews hundreds of pages of invention disclosures or prior art references. If AI-assisted review could materially reduce the risk of missing relevant documents, refusing to use an AI tool puts the attorney at risk of a malpractice suit.

A fourth example of AI non-use potentially constituting malpractice is based on the prohibition of attorneys charging an “unreasonable fee” under Model Rule 1.5[18] and 37 C.F.R. § 11.105. If use of an AI tool reduces the time needed to complete a task at the same quality level, even after accounting for necessary human verification, billing for unassisted manual work may be deemed unreasonable.

II. How can existing usage guidelines and standards be made clearer?

Multiple states now offer more explicit guidance beyond the directive in Comment 8 to “keep abreast of changes in the law and its practice.” For example, the Illinois ARDC’s “Illinois Attorney’s Guide to Implementing AI”[19] and the Illinois Supreme Court’s recent policy announcement[20] offer the beginnings of practical direction on how to choose appropriate generative AI tools, manage client rights, and implement internal policies. However, such jurisdictional guidance can fail to explicitly define competent AI practice. The mandate under Comment 8 to “staying abreast” of technological changes lacks clear operational boundaries, particularly alongside strict international frameworks. For example, the European Union’s AI Act identifies high-risk applications in judicial administration under Annex III,[21] and its track record of massive extraterritorial privacy enforcement under the GDPR[22] means any firm with European clients or ties should ensure compliant AI usage. To bridge this gap, we propose augmenting the Illinois ARDC’s guidelines with five firm-level rules to help attorneys stay ahead of the curve:

1. “Sandboxed” AI Platforms. Firms should designate approved AI tools that operate within secure, enterprise-grade environments and prohibit the use of public AI platforms for client matters.[23] These tools contractually guarantee confidentiality, restrict vendor prompt retention, prohibit training public models on client data, and provide robust audit capabilities.[24] For intellectual property practitioners handling technical disclosures, sandbox policies must mandate U.S.-based server nodes to ensure cloud endpoint routing does not trigger export control violations. Further, restricting attorneys to vetted, sandboxed AI environments significantly reduce confidentiality risks that may violate Model Rule 1.6[25] and 37 C.F.R. § 11.106.
2. Standardized Prompt Templates. Firms should develop centralized prompt libraries for standard tasks. These templates promote output consistency, reduce the effects of user bias[26], and reduce the time spent optimizing AI inputs for similar tasks.
3. Matter-Type Restrictions. AI tools are not appropriate for every matter. Firms should establish clear policies defining matters that require heightened approval or a total ban on AI tool use.[27] Firms should avoid entering client data into any AI tools (even sandboxed ones) if restricted by the EU AI Act or GDPR.
4. Documentation and Record-keeping. Firms should track material AI use, such as prompts, generated outputs, attorney edits, and final work product. This record-keeping aids internal performance reviews, satisfies client inquiries, defends against malpractice claims, and helps comply with court orders mandating AI disclosures and certifications[28].
5. Professional Liability Insurance Review. Firms should confirm that malpractice and liability insurance policies explicitly cover AI-assisted legal services, including research, drafting, document review, and client-facing AI applications like chatbots.[29] Where coverage gaps exist, firms should find supplemental protection before expanding AI integration.[30]

III. Conclusion

AI is here to stay and maintaining technological competence is essential. Federal judges[31] and bar associations[32] agree that avoiding AI may soon constitute malpractice. To ethically use AI, firms must establish internal standards, ensure client transparency, and actively engage with emerging regulatory frameworks. Rather than waiting for adverse court rulings, attorneys who proactively implement ethical AI practices today will shape AI standards for tomorrow.

[1] Mata v. Avianca, Inc., 678 F.Supp.3d 443, 464-65 (S.D.N.Y. 2023) (holding failure to cite check fictitious cases produced using generative AI was a violation of Rule 11).

[2] 60 F.2d 737 (2d Cir. 1932).

[3] Id. at 737-39.

[4] Id.

[5] Id. at 739-40.

[6] Id. at 740.

[7] Model Rules of Prof. Conduct r. 1.1 (A.B.A. 2026).

[8] Model Rules of Prof. Conduct r. 1.1 cmt. 8 (A.B.A. 2026)

[9] Forty states have adopted comment 8 in some form. See Litigation Technology Competence State Law Survey, LexisNexis Practical Guidance (2023), https://www.lexisnexis.com/pdf/practical-guidance/ai/litigation-technology-competence-state-law-survey.pdf.

[10] James v. Nat’l Fin. LLC, No. CV 8931-VCL, 2014 WL 6845560 (Del. Ch. Dec. 5, 2014).

[11] Id. at *25.

[12] Id.

[13] Id. at *26.

[14] Model Rules of Prof. Conduct r. 1.2 (A.B.A. 2026).

[15] Roy Strom, Clients Push Big Law Firms to Use Generative AI for Cost Savings, Bloomberg Law (Sept. 11, 2025, at 04:00 AM CDT), https://news.bloomberglaw.com/business-and-practice/clients-push-big-law-firms-to-use-generative-ai-for-cost-savings.

[16] Model Rules of Prof. Conduct r. 1.3 (A.B.A. 2026).

[17] How AI Enhances Legal Document Review, ABA: Technology Today (Feb. 13, 2025), https://www.americanbar.org/groups/law_practice/resources/law-technology-today/2025/how-ai-enhances-legal-document-review/.

[18] Model Rules of Prof. Conduct r. 1.5(a) (A.B.A. 2026).

[19] Illinois Attorney’s Guide to Implementing AI, Ill. Att’y Reg. & Disciplinary Comm’n (2025), https://iardc.org/Files/Implementing-AI-Guide/

[20] Illinois Supreme Court Policy on Artificial Intelligence, Ill. Sup.Ct. (2025), https://ilcourtsaudio.blob.core.windows.net/antilles-resources/resources/e43964ab-8874-4b7a-be4e-63af019cb6f7/Illinois%20Supreme%20Court%20AI%20Policy.pdf.

[21] Annex III: High-Risk AI Systems Referred to in Article 6(2), Artificial Intelligence Act, (Regulation (EU) 2024/1689).

[22] 20 Biggest GDPR fines so far [2025], Legit: Blog (Mar. 2025), https://legit.eu/resources/blog/5-biggest-gdpr-fines-so-far-2025/.

[23] See Illinois Attorney’s Guide to Implementing AI, supra note 16.

[24] See Michelle McSweeney, Cloud-Based Law Firm Software Vs. On-Premise Servers, Cilo Blog (Feb. 27, 2026), https://www.clio.com/uk/blog/cloud-based-law-firm-software-v-on-premise-servers/.

[25] Model Rules of Prof. Conduct r. 1.6 (A.B.A. 2026).

[26] Grace Chang & Heidi Grant, When AI Amplifies the Biases of Its Users, Harv. Bus. Rev. (Jan. 23, 2026), https://hbr.org/2026/01/when-ai-amplifies-the-biases-of-its-users. (Warning that relying on unreviewed AI outputs risks compounding user bias and injecting discriminatory or one-sided reasoning into legal analysis because models tend to mirror the user’s framing).

[27] See The Illinois Attorney’s Guide to Implementing AI, Table 3 (classifying data into general information, de-identified information, confidential information, sensitive personal information, and system-wide processing, while proposing client communication strategies tailored to each category).

[28] Briana Martinotti, Kaityln Stone & Michael C. Zogby, What to Know about Court-Mandated Disclosure of Artificial Intelligence in Court Submissions, ABA Litig. (Oct. 17, 2025), https://www.americanbar.org/groups/litigation/resources/newsletters/mass-torts/court-mandated-disclosure-artificial-intelligence-court-submissions/.

[29] Standard legal malpractice policies may not cover AI-related errors. See Danielle Braff, Does your professional liability insurance cover AI mistakes? Don’t be so sure, ABA J. (Feb. 1, 2025), https://www.americanbar.org/groups/journal/articles/2025/does-your-professional-liability-insurance-cover-ai-mistakes-dont-be-so-sure/?abajoin=true.

[30] Rod Bolivar, AI claims reach legal malpractice market, Ins. Bus. Mag. (May 26, 2026), https://www.insurancebusinessmag.com/us/news/professional-liability/ai-claims-reach-legal-malpractice-market-576535.aspx.

[31] Mike Vilensky, NY Federal Judge Questions if Avoiding AI Could Be Malpractice, Bloomberg Law (Jan. 13, 2026, at 04:00 PM CST),
https ://news.bloomberglaw.com/new-york-brief/ny-federal-judge-questions-if-avoiding-ai-could-be-malpractice. (“I heard somebody say employers are risking malpractice by relying too much on AI . . . I think there may come a point where it’s the opposite—where you’re committing malpractice if you don’t incorporate AI into your practice.”).

[32] Patrick T. Barone, Is saying no to AI saying yes to legal malpractice? Our profession’s new dilemma, Mich. Bar. J. (Jan. 2026), https://www.michbar.org/journal/Details/Is-saying-no-to-AI-saying-yes-to-legal-malpractice-Our-professions-new-dilemma?ArticleID=5219.