Ethical Duty of Technological Competence

Most attorneys understand that their ethical duty of competence requires having the knowledge, skill, and preparation to handle a particular matter, whatever the complexities of that matter.  

However, many attorneys have struggled with this long-established ethical principle in the face of rapid technological change impacting litigation practice.  Some attorneys are ignorant of evolving technologies or fear them as unduly complex, while others rely too much on technologies and fail to understand their limitations.  For many litigators, e-discovery has remained terrain “where angels fear to tread” (United States v. O’Keefe, 537 F. Supp. 2d 14, 24 (D.D.C. 2008) (Facciola, Mag. J.)).

Technological issues are no longer relegated to substantial cases or large corporate matters.  Instead, they pervade all litigation, regardless of size or complexity, particularly in a post-COVID-19 world.

For example, most clients regularly communicate by email, text message, instant message, and ephemeral message (whether through collaboration platforms or third-party applications), and they actively use some form of social media.  Additionally, counsel routinely handle depositions, appear at hearings, and conduct trials using collaborative video technology.  Due to the explosion of electronic platforms and increasing reliance on ever-changing technology as a tool in virtually every aspect of litigation, technological ignorance has long ceased to be an excuse for counsel.  The rise of GenAI underscores this reality and the need for technological competence.  (For more information, see Key Legal Issues in Using Generative AI: Overview (US) on Practical Law.)  

Rules and Guidance

Most bar associations, state bars, and courts have issued at least some guidance on counsel’s duty of competence in the e-discovery context.  Notable advice can be found in the following sources, which are described in more detail in this article:

ABA Model Rule 1.1

Model Rule 1.1 requires attorneys to provide competent representation to a client.  This requires “the legal knowledge, skill, thoroughness and preparation reasonably necessary for the representation” (Model Rule 1.1).  Comment 8 to this rule clarifies that, as part of this duty, attorneys must keep abreast of changes in the law and its practice, including the “benefits and risks associated with relevant technology.”

The ABA stated in the 2012 committee report accompanying the amendment to comment 8 that technology is an integral aspect of law practice and the amendment does not impose any new obligations.  It explained that, instead, “the amendment is intended to serve as a reminder to lawyers that they should remain aware of technology, including the benefits and risks associated with it, as part of a lawyer’s general ethical duty to remain competent.”  (ABA Commission on Ethics 20/20, Resolution & Report: Technology & Confidentiality, at 3 (Aug. 2012).)

As later opinions clarify, attorneys cannot abdicate their responsibility to understand technology or delegate that responsibility entirely to IT departments, outside vendors, paralegals, and secretaries.  As technology evolves, particularly in the area of cybersecurity, attorneys have more issues and risks to monitor than ever before.  (See, for example, ABA Formal Op. 483 (Oct. 17, 2018) (stating that attorneys have a duty to monitor for and respond to data breaches); ABA Formal Op. 477R (May 22, 2017) (discussing attorneys’ obligation to secure the communication of protected client information).)  The duty to monitor includes providing supervision and oversight over the use of GenAI.  

On July 29, 2024, to address the rapidly growing use of AI in law firms, the ABA issued Formal Opinion 512, which is its first ethics guidance on the use of GenAI in legal practice consistent with the Model Rules.  The opinion highlights six key ethical duties that GenAI implicates:

With respect to Model Rule 1.1, the opinion confirms that attorneys must either:

Although attorneys are not required to become GenAI experts, they must recognize the potential for error and ensure the GenAI tool’s output is accurate and appropriate for the intended task.  This includes critically reviewing and independently verifying AI-generated content before relying on it in client work or court filings.  

Formal Opinion 512 acknowledges that the appropriate amount of review and verification depends on the GenAI tool used and the specific task it performs.  However, the opinion warns that attorneys may not abdicate their responsibilities by relying solely on GenAI tools to perform tasks that call for professional judgment.  Moreover, because GenAI tools are evolving rapidly, attorneys must stay informed, such as through continuing legal education (CLE) courses, reading materials targeted at the legal profession, and consulting with knowledgeable colleagues or AI experts.  Attorneys who fail to follow this guidance risk significant consequences, including sanctions and reputational harm.

State Bar Guidance

Most state bar organizations have provided some form of guidance on attorneys’ duty of technological competence.  They generally follow the ABA’s approach, often adopting Model Rule 1.1 verbatim (see, for example, Del. Lawyers’ Rules of Prof’l Conduct, R. 1.1, cmt. 8; Mass. Rules of Prof’l Conduct, R. 1.1, cmt. 8 (using nearly verbatim language)).

According to the ABA’s list of jurisdictions, as of April 2023, 39 states had adopted some form of statement about technological competence, while another six states had adopted comments to their Rule 1.1 but no statement on technological competence.  The District of Columbia recently amended the comments to its Rule of Professional Conduct 1.1 with similar language to comment 8 of ABA Model Rule 1.1 (D.C. Court of Appeals, No. M284-24 (April 7, 2025)).  Similarly, Puerto Rico has adopted new Rule of Professional Conduct 1.19 (effective January 1, 2026), which requires technological competence and diligence and goes beyond the ABA Model Rules (Puerto Rico Supreme Court Order, 2025 WL 1797628 (June 17, 2025)).  

Further, many states provide specific additional guidance defining competence with technology.  For example, in 2015, the State Bar of California issued an extremely influential formal opinion (which continues to be routinely cited) holding that an attorney is not competent to handle complex cases involving ESI without having sufficient understanding of the technical skills, knowledge, and aptitude required to conduct e-discovery (see Cal. State Bar Standing Comm. on Prof’l Resp. & Conduct, Formal Op. No. 2015-193 (June 30, 2015)).  

The opinion states that attorneys undertaking complex litigation involving e-discovery should be able to perform at least nine specific tasks, including: 

The opinion directs that attorneys who handle litigation may not simply ignore the potential impact of evidentiary information existing in electronic form.  Attorneys who are not sufficiently competent in these areas must either:

The opinion states that a lack of competency in e-discovery matters can, in certain circumstances, result in ethical violations (Cal. State Bar Formal Op. No. 2015-193, at 3, 7 (citing Rule 3-110(C), which states that “a lack of technological knowledge in handling e-discovery may render an attorney ethically incompetent to handle certain litigation matters involving e-discovery, absent curative assistance … even where the attorney may otherwise be highly experienced”)).

Recent state bar guidance also addresses the use of GenAI tools similarly to the ABA’s Formal Opinion 512.  The guidance cautions attorneys to verify AI-generated content and ensure compliance with ethical obligations, and emphasizes the need for supervision, accuracy, and the protection of client confidentiality.  For example, the Pennsylvania and Philadelphia Bar Associations jointly issued Formal Opinion 2024-200, titled Ethical Issues Regarding the Use of Artificial Intelligence (May 22, 2024).  This opinion warns attorneys against relying on GenAI outputs without independent verification and emphasizes the need for technological competence and compliance with the following Pennsylvania Rules of Professional Conduct:

Similar opinions from other jurisdictions include:

Moreover, the number of jurisdictions issuing and updating GenAI guidance continues to increase rapidly.  The New York City Bar created a May 2025 summary of ethics opinions and reports related to GenAI that compiled formal ethics opinions from California, the District of Columbia, Florida, Kentucky, Michigan, North Carolina, New York, Pennsylvania, Minnesota, Missouri, West Virginia, the Massachusetts Attorney General, and the New York Attorney General.  

2015 Amendments to the FRCP and Relevant Case Law

The 2015 amendments to the FRCP significantly impacted e-discovery by, for example, limiting the scope of discovery, clarifying the standards for the preservation of ESI, and streamlining the discovery process (for more information, see Overview of December 2015 Amendments to the Federal Rules of Civil Procedure on Practical Law).  

While the proportionality standard in amended FRCP 26 has rightfully garnered significant litigation attention, amended FRCP 37(e) on sanctions for lost ESI has also been extremely influential, particularly as new forms of ESI have become increasingly commonplace.  FRCP 37(e) requires counsel to ensure that their clients take reasonable steps to preserve ESI and provides that sanctions can be entered without a finding of prejudice if the failure to take reasonable steps was intentional.

Consider, for example, text messages that are as ubiquitous today as emails were a decade ago.  Numerous federal cases throughout the country have weighed in on the need to preserve text messages, and virtually all do so under the framework of FRCP 37(e) (see Colonies Partners, L.P. v. Cnty. of San Bernardino, 2020 WL 1496444 (C.D. Cal. Feb. 27, 2020), report and recommendation adopted, 2020 WL 1491339 (C.D. Cal. Mar. 27, 2020) (imposing sanctions under FRCP 37(e) for spoliation of text messages); Paisley Park Enters., Inc. v. Boxill, 330 F.R.D. 226 (D. Minn. 2019) (imposing sanctions under FRCP 37(b) and 37(e) after rejecting the claim that the cell phones at issue were personal devices given their business use); Sinclair v. Cambria Cnty., 2018 WL 4689111 (W.D. Pa. Sept. 28, 2018) (imposing sanctions for text message spoliation under Rule 37(e)); First Fin. Sec., Inc. v. Freedom Equity Grp., LLC, 2016 WL 5870218 (N.D. Cal. Oct. 7, 2016) (same)).  

Under FRCP 37(e), even a failure to disable auto-delete settings can subject the client (and counsel) to sanctions (see Mod. Remodeling, Inc. v. Tripod Holdings, LLC, 2021 WL 3852323, at *10 (D. Md. Aug. 27, 2021) (imposing sanctions under FRCP 37(e), explaining that the “[f]ailure to either disengage the auto-delete setting or to back up messages to a cloud server prior to deleting them from a device ‘is sufficient to show’ that a defendant ‘acted unreasonably’”)).  FRCP 37(e) applies to all forms of ESI, such as information contained on social media (see Torgersen v. Siemens Bldg. Tech., Inc., 2021 WL 2072151 (N.D. Ill. May 24, 2021) (imposing sanctions under FRCP 37(e) for the intentional deletion of a Facebook page)).

To avoid sanctions, it is simply not enough to tell a client to preserve everything.  Counsel must take further steps, including but not limited to:

GenAI Hallucination Cases and Judicial Responses

As set out above, the rise of GenAI tools, whether general purpose (for example, ChatGPT, Gemini, and Copilot) or legal focused (such as CoCounsel and HarveyAI), has introduced new ethical and practical challenges for attorneys.  While these tools can enhance efficiency when researching, drafting, and communicating with clients, they also pose profound risks related to accuracy, confidentiality, and candor to the tribunal.

Indeed, it is well established that the current iteration of GenAI tools are prone to invent content in response to user queries and are more likely to fabricate information than return no information at all.  Despite appearances, GenAI tools do not think at all, much less think critically.  As the ABA’s Formal Opinion 512 explains, “G[en]AI tools lack the ability to understand the meaning of the text they generate or evaluate its context. Thus, they may combine otherwise accurate information in unexpected ways to yield false or inaccurate results.  Some G[en]AI tools are also prone to ‘hallucinations,’ providing ostensibly plausible responses that have no basis in fact or reality.”

Numerous cases throughout the country emphasize the dire nature of these risks, with attorneys facing sanctions and disciplinary actions at an alarming rate, almost invariably for citing non-existent facts, cases, or other materials.  In one of the earliest (and accordingly most cited and discussed) hallucination cases, Mata v. Avianca, Inc. (678 F. Supp. 3d 443 (S.D.N.Y. 2023)), the court sanctioned attorneys who submitted a brief containing fictitious case citations generated by ChatGPT.  The court imposed a $5,000 fine and ordered the attorneys to write letters to their client and the judges falsely identified as authors of the fake opinions, informing them of their actions.  The court emphasized the duty to verify all legal authorities and warned against blind reliance on AI tools, setting out several harms flowing from the submission of fake opinions.

Similarly, in Lacey v. State Farm General Insurance Co., the court criticized the use of AI-generated content in filings without proper review, striking the offending briefs and awarding $31,100 to the opposing party (2025 WL 1363069 (C.D. Cal. May 5, 2025)).  In Park v. Kim, the Second Circuit referred the responsible attorney to the court’s grievance panel for further investigation after discovering that half of the cases the attorney cited in the appellate reply brief were fabricated (91 F.4th 610 (2d Cir. 2024)). In extreme cases, sanctions can extend beyond the attorney, with at least one court dismissing multiple matters for repeat violations (ByoPlanet Int’l, LLC v. Johansson, 2025 WL 2091025 (S.D. Fla. July 17, 2025)).  

Hallucination cases have become so commonplace that the court in United States v. Hayes provided a list of nine of these cases within the short span of one paragraph (763 F. Supp. 3d 1054, 1071-72 (E.D. Cal. 2025), reconsideration denied, 2025 WL 1067323 (E.D. Cal. Apr. 9, 2025)).  Even attorneys at sophisticated law firms with established policies and AI committees are not immune (see, for example, as Johnson v. Dunn, 2025 WL 2086116 (N.D. Ala. July 23, 2025)).  Perhaps more alarming is that these are but a handful of the decisions that have drenched the legal landscape since ChatGPT’s November 2022 release opened the floodgates of GenAI.  In fact, one researcher’s website devoted to tracking GenAI hallucination cases listed 545 of these cases as of November 14, 2025.

In response to this proliferation of erroneous AI-generated information, many judges have begun implementing safeguards, such as a requirement for attorneys to certify that no part of a document was generated by an AI tool or that a human reviewed any AI-generated text in briefs or other court filings.  

While most judges who have addressed GenAI have instituted these disclaimer requirements, there are a few who have entirely banned the use of GenAI tools (see, for example, Judge Christopher A. Boyko’s standing order on the use of GenAI in the Northern District of Ohio (providing that “no attorney for a party, or a pro se party, may use Artificial Intelligence (‘AI’) in the preparation of any filing submitted to the Court.  Parties and their counsel who violate this AI ban may face sanctions including, inter alia, striking the pleading from the record, the imposition of economic sanctions or contempt, and dismissal of the lawsuit”)).  Responsible AI in Legal Services provides a useful table that tracks and summarizes judicial standing orders and local court rules on the use of AI.  

Best Practices

Now that counsel’s e-discovery responsibilities have been the subject of significant consideration and clarification, counsel must face the continuing challenge of managing their e-discovery efforts while handling evolving technology like GenAI.  They must adopt tangible best practices to ensure compliance with their ethical obligations.

In the e-discovery context, clients are increasingly holding attorneys responsible for failing to meet their duty of technological competence (see, for example, Indus. Quick Search, Inc. v. Miller, Rosado & Algois, LLP, 2018 WL 264111 (S.D.N.Y. Jan. 2, 2018) (allowing certain malpractice claims alleging negligence in e-discovery preservation to move forward)).  Discovery on discovery disputes are also on the rise.  These disputes examine a party’s collection, retrieval, and production efforts and are costly in terms of time, money, and the reputation of both the client and their counsel.  (For more on the issues courts consider when determining when and how a party may obtain information about another party’s preservation and discovery processes, see Discovery on Discovery (Federal) on Practical Law.) 

In light of the guidance defining technological competence (such as the nine subject matters enumerated by the State Bar of California), all attorneys involved in e-discovery must critically ask themselves whether they have the skills, knowledge, and ability to:

Beyond e-discovery, attorneys must have the skills, knowledge, and ability to:

In any given case, counsel should at a minimum observe the following ethical practices in litigation:

(For more information, see Avoiding Common Ethical Pitfalls in E-Discovery and Generative AI Ethics for Litigators on Practical Law.)