Artificial Intelligence and Professional Conduct

Considering the Ethical Implications of Using Electronic Legal Assistants

January/February 2024 By Maria Berkenkotter and Lino Lipinsky de Orlov

The ethical rules that apply to lawyers and judges are meant to evolve as society changes. This article poses questions that should be considered when adapting these rules to reflect the increased use of artificial intelligence tools in the legal profession.

The creatures outside looked from robot to man, and from man to robot, and from robot to man again; but already it was impossible to say which was which.1

The revolution in artificial intelligence (AI) has inspired commentators, lawyers, and judges to consider the implications of these new technological capabilities on the practice of law and the courts. Although AI has aided practitioners and the judiciary since LexisNexis and Westlaw unveiled their computerized research services in the 1970s, ChatGPT and similar generative AI tools that simulate the text of human authors have led to conversations about the role of and risks associated with automated writers. These AI resources, while impressive in their ability to churn out poetry or prose in seconds, are not yet capable of competently duplicating the work of judges, lawyers, and other legal professionals. The popular press has regaled readers with stories about lawyers who lacked a basic understanding of the limitations of AI and filed ChatGPT-drafted motions and briefs filled with impressive but fictitious legal citations.

Given the recent leaps in the capabilities of generative AI, the legal profession cannot wait to consider the implications of robot authors for the ethics rules applicable to lawyers and judges, as well as the rules governing the unauthorized practice of law.

This article considers which of those rules may require amendment in a world in which lawyers turn to generative AI platforms to draft their motions, briefs, and memorandums, and in which potential clients or self-represented litigants may communicate with a chatbot before speaking with a human lawyer or a court employee. Specifically, it identifies several provisions within the Colorado Rules of Professional Conduct, the Colorado Code of Judicial Conduct, and the Colorado Unauthorized Practice of Law (UPL) Rules2 that the use of generative AI may implicate.

We hope to foster a discussion in the Colorado legal community rather than propose specific changes to these ethical standards. Late last June, the Colorado Supreme Court asked the Standing Committee on the Colorado Rules of Professional Conduct to form a subcommittee to consider recommendations for amendments to those rules to address lawyers’ use of AI tools. In addition, the Court is examining whether changes are needed to the Colorado Code of Judicial Conduct and the Colorado UPL Rules to respond to the legal profession’s increasing use of AI.

Legal writers with a greater knowledge and understanding of the technological aspects of AI than we possess have published articles explaining how these tools were developed, how they operate, and their limitations.3 We lack the technological expertise—and space in this article—to provide a primer on these important topics. We urge lawyers and judges to educate themselves about the basics of AI as our profession explores the implications of this new leap in technology.

Colorado Rules of Professional Conduct

The drafters of the American Bar Association (ABA) Model Rules of Professional Conduct and the earlier ethical codes for lawyers intended for these standards to evolve to address unanticipated changes in society and the law.4 For example, in 2012, in response to the “sometimes bewildering pace of technological change,” the ABA amended comment 8 to Rule 1.1 of the Model Rules, which concerns the duty of competence, to add a reference to lawyers’ technological competence.5 The Colorado Supreme Court approved a similar change to the analogous comment in the Colorado Rules of Professional Conduct.

We believe that reexamination of these rules is warranted in light of the impact of the widespread use of generative AI platforms, such as ChatGPT, on the practice of law. Today’s lawyers can only benefit from guidance regarding the risks and potential consequences of using these tools. To assist in this analysis, we have highlighted those provisions that may be implicated by the use of generative AI, followed by questions that will likely arise as lawyers, nonlawyers, and judges increasingly use this technology. We do not offer opinions on which provisions should be amended or what those amendments might look like. We leave those issues to the Standing Committee on the Colorado Rules of Professional Conduct and the Supreme Court.

The Duty of Competence

Colo. RPC 1.1 provides:

A lawyer shall provide competent representation to a client. Competent representation requires the legal knowledge, skill, thoroughness and preparation reasonably necessary for the representation.

Further, comment 8 to this rule states:

To maintain the requisite knowledge and skill, a lawyer should keep abreast of changes in the law and its practice, and changes in communications and other relevant technologies . . . .

Considerations: As generative AI use increases, is the general reference to “technologies” in comment 8 sufficient to alert lawyers to the risks and limitations of these new electronic tools? For example, not all lawyers may know that, at least as of early 2024, generative AI platforms are incapable of conducting legal research, reliably analyzing legal issues, or checking the completeness or accuracy of legal writing. The popular press has reported, with more than a modicum of glee, on lawyers who relied on ChatGPT to draft motions, only to learn to their horror from the court or opposing counsel that the legal citations in their filing were fictitious.6 As part of their professional duties, do lawyers need to possess a basic understanding of how AI resources—and, in particular, generative AI—function and what they can and cannot accomplish?7

The Duty to Communicate With Clients

Colo. RPC 1.4 provides:

(a) A lawyer shall:

(1) promptly inform the client of any decision or circumstance with respect to which the client’s informed consent, as defined in Rule 1.0(e), is required by the Rules of Professional Conduct;

(2) reasonably consult with the client about the means by which the client’s objectives are to be accomplished;

. . . .

(b) A lawyer shall explain a matter to the extent reasonably necessary to permit the client to make informed decisions regarding the representation.

Considerations: Does a lawyer need to obtain informed consent from a client before using an AI tool to draft a document on the client’s behalf, considering the current stage of the development of generative AI? Does the client need to be apprised that the lawyer intends to rely on an automated resource, rather than on a trained legal professional, for the first and possibly subsequent drafts of certain documents? Does the lawyer need to inform the client about the potential risks of such a practice? For example, should a lawyer intending to generate drafts using an AI platform explain to the client the procedures the lawyer has put in place to edit, review the accuracy of computer-generated text, and eliminate possible bias in that work product? Further, does a lawyer need to consider the risks of not employing generative AI when preparing preliminary drafts for a client? Does a lawyer need to inform a client of the potential cost of not using AI when some tasks—particularly ones involving repetitive labor and completion of simple forms—may be accomplished more efficiently, and at a lower cost to the client, by using AI rather than a human writer?

Reasonableness of Fees

Colo. RPC 1.5 provides:

(a) A lawyer shall not make an agreement for, charge, or collect an unreasonable fee or an unreasonable amount for expenses. The factors to be considered in determining the reasonableness of a fee include the following:

(1) the time and labor required, the novelty and difficulty of the questions involved, and the skill requisite to perform the legal service properly;

. . . .

(3) the fee customarily charged in the locality for similar legal services;

(4) the amount involved and the results obtained;

(5) the time limitations imposed by the client or by the circumstances;

. . . .

(7) the experience, reputation, and ability of the lawyer or lawyers performing the services; and

(8) whether the fee is fixed or contingent.

Considerations: How will clients benefit financially from the efficiencies created through a lawyer’s use of generative AI? How will a lawyer decide what is a reasonable fee for drafting a document using AI, as well as a reasonable fee for creating a document that could have been prepared using AI to produce the initial draft? Should a lawyer in private practice consider the use or non-use of generative AI in setting the lawyer’s standard rates? Should a judge consider a lawyer’s use or non-use of generative AI in determining the reasonableness of requested attorney fees?

Confidentiality of Information

Colo. RPC 1.6 provides:

(a) A lawyer shall not reveal information relating to the representation of a client unless the client gives informed consent . . . .

. . . .

(c) A lawyer shall make reasonable efforts to prevent the inadvertent or unauthorized disclosure of, or unauthorized access to, information relating to the representation of a client.

► Considerations: Does a lawyer violate Colo. RPC 1.6 by providing a client’s confidential information to a third-party generative AI vendor without the client’s knowledge or consent? What safeguards must a lawyer put in place to protect confidential client information when drafting generative AI prompts and queries? As with the use of other resources that involve the provision of data to a third-party vendor, such as tools allowing for storage of client documents, attorney work product, and other confidential information in the cloud, what steps must a lawyer take to satisfy the duty of confidentiality when employing an AI platform? What type of reasonable security precautions must a lawyer take to protect a client’s data from inadvertent disclosure? Similarly, does a lawyer violate the duty of confidentiality by submitting a generative AI query that includes a client’s confidential information, given that unauthorized persons may be able to access such information?

Candor to the Tribunal

Colo. RPC 3.3 provides:

(a) A lawyer shall not knowingly:

(1) make a false statement of material fact or law to a tribunal or fail to correct a false statement of material fact or law previously made to the tribunal by the lawyer . . . .

Considerations: Given that no generative AI tool currently possesses the ability to (1) draft an accurate and comprehensive legal analysis containing correct legal citations, (2) apply legal principles to novel facts, or (3) exercise professional judgment, what steps should lawyers take to satisfy the duty of candor to the court when the lawyer uses generative AI? Is it prudent for a lawyer to assume that the tool has generated only a rough first draft that requires a complete review and thorough editing? (And, if so, would using such a tool result in saving lawyers time and resources?) Moreover, does the duty of candor regarding use of generative AI only extend to lawyers’ appearances in courts that have adopted a practice standard or entered a standing order requiring lawyers to disclose whether they employed a generative AI tool to draft motions, briefs, or other documents submitted to the court, or should lawyers automatically disclose such use?8

Responsibilities of a Partner or Supervisory Lawyer and Responsibilities Regarding Nonlawyer Assistants

Colo. RPC 5.1 provides:

(a) A partner in a law firm,[9] and a lawyer who individually or together with other lawyers possesses comparable managerial authority in a law firm, shall make reasonable efforts to ensure that the firm has in effect measures giving reasonable assurance that all lawyers in the firm conform to the Rules of Professional Conduct.

(b) A lawyer having direct supervisory authority over another lawyer shall make reasonable efforts to ensure that the other lawyer conforms to the Rules of Professional Conduct.

(c) A lawyer shall be responsible for another lawyer’s violation of the Rules of Professional Conduct  if:

(1) the lawyer orders or, with knowledge of the specific conduct, ratifies the conduct involved;

(2) the lawyer is a partner or has comparable managerial authority in the law firm in which the other lawyer practices, or has direct supervisory authority over the other lawyer, and knows of the conduct at a time when its consequences can be avoided or mitigated but fails to take reasonable remedial action.

Colo. RPC 5.3 provides:

With respect to nonlawyers employed or retained by or associated with a lawyer:

(a) a partner, and a lawyer who individually or together with other lawyers possesses comparable managerial authority in a law firm shall make reasonable efforts to ensure that the firm has in effect measures giving reasonable assurance that the person’s conduct is compatible with the professional obligations of the lawyer;

(b) a lawyer having direct supervisory authority over the nonlawyer shall make reasonable efforts to ensure that the person’s conduct is compatible with the professional obligations of the lawyer; and

(c) a lawyer shall be responsible for conduct of such a person that would be a violation of the Rules of Professional Conduct if engaged in by a lawyer if:

(1) the lawyer orders or, with the knowledge of the specific conduct, ratifies the conduct involved; or

(2) the lawyer is a partner or has comparable managerial authority in the law firm in which the person is employed, or has direct supervisory authority over the person, and knows of the conduct at a time when its consequences can be avoided or mitigated but fails to take reasonable remedial action.

Considerations: How can a lawyer ensure that the lawyers and nonlawyers whom the lawyer supervises are aware of the risks and limitations of generative AI tools? What type of training does the lawyer need to provide to ensure that team members are properly trained in using AI resources? Must such training be supplemented with every rollout of an AI tool that offers new features, new capabilities, or new risks? Given the rapid pace of innovation, how often should the lawyer provide the training? In addition, what guidance should a lawyer provide to those the lawyer supervises regarding using generative AI at work? Does a lawyer violate Colo. RPC 5.1 by not requiring those under the lawyer’s supervision to disclose their use of generative AI on client work? What happens if the supervised lawyer uses an AI tool that provides inaccurate results, but the supervisory lawyer submits the AI-generated filing without identifying the AI component or the inaccuracies? Further, at this stage in the development of generative AI, is it reasonable for a lawyer to prohibit team members from using such a platform when preparing documents for a client? How is using generative AI to draft a legal document, followed by careful cite-checking and editing, any different from relying on a first-year law student intern to prepare such a document?

Conduct Involving Dishonesty, Fraud, Deceit, or Misrepresentation

Colo. RPC 8.4(c) provides:

It is professional misconduct for a lawyer to:

. . . .

(c) engage in conduct involving dishonesty, fraud, deceit or misrepresentation . . . .

Considerations: The principle underlying Colo. RPC 8.4(c) is similar to that underlying CRCP 11(a), which states that, by signing a pleading, a lawyer certifies, among other representations, that the lawyer read the pleading and, to the best of the lawyer’s knowledge, information, and belief formed after reasonable inquiry, it is well grounded in fact and is warranted by existing law or a good faith argument for the extension, modification, or reversal of existing law. What type of reasonable inquiry must a lawyer undertake to confirm that an AI-generated pleading complies with CRCP 11(a)? What safeguards can a lawyer put in place when drafting and editing a pleading that includes some AI-generated text? What safeguards are necessary and appropriate when a lawyer uses an AI platform to review electronic discovery materials? And does a lawyer violate Colo. RPC 8.4(c) by failing to disclose to opposing counsel or a court that the lawyer employed AI to enhance or otherwise edit a photograph or graphical image submitted as an exhibit?

Conduct Prejudicial to the Administration of Justice

Colo. RPC 8.4(d) provides:

It is professional misconduct for a lawyer to:

. . . .

(d) engage in conduct that is prejudicial to the administration of justice . . . .

Considerations: Would a lawyer violate this rule by submitting an AI-generated document to a client or to a court if the document contains false factual or legal statements that the lawyer failed to identify through a reasonable review? Is a lawyer’s known use of generative AI in and of itself prejudicial to the administration of justice, to the extent it creates the perception that the lawyer did not use his or her skills and training in representing the client but, rather, delegated the lawyer’s legal abilities to a machine? Or does the use of generative AI suggest that a machine is capable of replacing the skills and training inherent in hiring a legal professional?

Bias

Colo. RPC 8.4(g) provides:

It is professional misconduct for a lawyer to:

. . . .

(g) engage in conduct, in the representation of a client, that exhibits or is intended to appeal to or engender bias against a person on account of that person’s race, gender, religion, national origin, disability, age, sexual orientation, or socioeconomic status, whether that conduct is directed to other counsel, court personnel, witnesses, parties, judges, judicial officers, or any persons involved in the legal process . . . .

Considerations: ChatGPT and other generative AI tools function by predicting the appropriate next word in text. They analyze the prior word or words before delivering the next word or words until they complete a full coherent sentence. For example, ChatGPT’s creators at OpenAI “taught” the tool by inputting vast amounts of written material from a variety of sources, such as newspaper articles, websites, and online postings. These inputs, however, include racist, sexist, and other biased content. Consequently, queries to ChatGPT can result in text containing biased and other offensive language. How can a lawyer safeguard against a generative AI platform’s implicit or explicit bias? Is a lawyer’s failure to detect this bias in an AI-generated pleading or document professional misconduct under Colo. RPC 8.4(g)?

Colorado Code of Judicial Conduct

We believe that examining the Colorado Code of Judicial Conduct is also warranted because the widespread use of generative AI will impact the work of the courts. As with the Rules of Professional Conduct, we do not offer opinions on which provisions of the Code of Judicial Conduct might merit amendment or what those amendments might look like. We leave those issues to the Supreme Court. Instead, we raise questions about issues that will likely arise as lawyers, nonlawyers, and judges increasingly use generative AI.

Promoting Confidence in the Judiciary

Colo. CJC 1.2 provides:

A judge shall act at all times in a manner that promotes public confidence in the independence, integrity, and impartiality of the judiciary, and shall avoid impropriety and the appearance of impropriety.

. . . .

Further, comment 1 to this rule states:

Public confidence in the judiciary is eroded by improper conduct and conduct that creates the appearance of impropriety.

Considerations: Could a judge’s use of generative AI erode public confidence in the integrity of the judiciary? For instance, would public knowledge that a judge uses generative AI as a drafting tool promote public confidence in the integrity of the judiciary, detract from it, or have no impact at all? If the judge carefully proofreads, cite-checks, and edits an order or decision created through generative AI, and the document completely and accurately reflects the judge’s ruling, would this scenario be different from a judge’s use of a computerized legal research tool such as Westlaw or LexisNexis? How would this analysis differ if the judge wasn’t diligent and didn’t understand generative AI’s shortcomings? What if the order included hallucinated citations and became the topic of a news story?10

Impartiality and Fairness and Bias, Prejudice, and Harassment

Colo. CJC 1.2 provides:

A judge shall uphold and apply the law, and shall perform all duties of judicial office fairly and impartially.

Colo. CJC 2.3 provides:

(A) A judge shall perform the duties of judicial office, including administrative duties, without bias or prejudice.

(B) A judge shall not, in the performance of judicial duties, by words or conduct manifest bias or prejudice, or engage in harassment, including but not limited to bias, prejudice, or harassment based upon race, sex, gender, religion, national origin, ethnicity, disability, age, sexual orientation, marital status, socioeconomic status, or political affiliation, and shall not permit court staff, court officials, or others subject to the judge’s direction and control to do so.

. . . .

(D) A judge shall require lawyers in proceedings before the court to refrain from manifesting bias or prejudice, or engaging in harassment, based upon attributes including but not limited to race, sex, gender, religion, national origin, ethnicity, disability, age, sexual orientation, marital status, socioeconomic status, or political affiliation, against parties, witnesses, lawyers, or others.

Considerations: As noted, queries to generative AI tools can result in racist, sexist, and other biased text due to the way large language model training works. How can a judge who uses generative AI ensure that the AI tool’s explicit or implicit bias does not infect the judge’s performance and undermine the judge’s impartiality? Given this known risk of bias, what, if anything, does Colo. CJC 2.3(D) require a judge to do if the judge knows a lawyer is using generative AI in connection with court proceedings?

Competence, Diligence, and Cooperation

Colo. CJC 2.5 provides:

(A) A judge shall perform judicial and administrative duties[] competently and diligently.

(B) A judge shall cooperate with other judges and court officials in the administration of court business.

Further, comment 1 to this rule states:

Competence in the performance of judicial duties requires the legal knowledge, skill, thoroughness, and preparation reasonably necessary to perform a judge’s responsibilities of judicial office.

Considerations: Like lawyers, judges have an obligation to be competent in performing their duties, which requires an up-to-date understanding of “legal knowledge, skill, thoroughness, and preparation.”11 Does this continuing competence requirement encompass having at least a basic understanding of how to use generative AI and an awareness of its known shortcomings, including hallucinated citations and the risk of bias? What steps must a judge take to learn about these issues? In addition, how would a judge’s misuse of generative AI impact the rights of the parties?12

Ensuring the Right to Be Heard

Colo. CJC 2.6 provides:

(A) A judge shall accord to every person who has a legal interest in a proceeding, or that person’s lawyer, the right to be heard according to law.

(B) A judge may encourage parties to a proceeding and their lawyers to settle matters in dispute but shall not act in a manner that coerces any party . . . .

Further, comment 2 to this rule states:

The steps that are permissible in ensuring a self-represented litigant’s right to be heard according to law include but are not limited to liberally construing pleadings; providing brief information about the proceeding and evidentiary and foundational requirements; modifying the traditional order of taking evidence; attempting to make legal concepts understandable; explaining the basis for a ruling; and making referrals to any resources available to assist the litigant in preparation of the case. Self-represented litigants are still required to comply with the same substantive law and procedural requirements as represented litigants.

Considerations: Should a judge alert self-represented litigants to the availability, benefits, and risks of generative AI resources? Similarly, if lawyers representing clients are using generative AI to create initial drafts of pleadings and other court filings, should or must a judge allow an unrepresented litigant to do the same? How does generative AI impact access to justice and the right to be heard, and what role should a judge play in ensuring that non-lawyers have access to generative AI as a tool that may enhance their right to be heard?

Supervisory Duties

Colo. CJC 2.12 provides:

(A) A judge shall require court staff, court officials, and others subject to the judge’s direction and control to act in a manner consistent with the judge’s obligations under this Code.

(B) A judge with supervisory authority for the performance of other judges shall take reasonable measures to ensure that those judges properly discharge their judicial responsibilities, including the prompt disposition of matters before them.

Considerations: The considerations discussed above in the context of Colo. RPC 5.1 also apply to judges through Colo. CJC 2.12. What is the impact of this provision on a judge’s duty to ensure that court staff only uses generative AI to the extent that the Code allows the judge to do so? As generative AI becomes more prevalent, could judges be faced with a need to establish parameters defining when court staff can and cannot use generative AI to assist with their official duties? Can a judge allow a law clerk to use generative AI to produce substantive first drafts of orders? Can a judge allow a staff member to use generative AI to assist in preparing non-substantive orders, like scheduling orders? Should a judge instruct court staff not to use generative AI for any official writing? What responsibilities does a chief judge have with respect to the use and oversight of generative AI?

Responding to Judicial and Lawyer Misconduct

Colo. CJC 2.15 provides:

(A) A judge having knowledge that another judge has committed a violation of this Code that raises a substantial question regarding the judge’s honesty, trustworthiness, or fitness as a judge in other respects shall inform the appropriate authority.

(B) A judge having knowledge that a lawyer has committed a violation of the Rules of Professional Conduct that raises a substantial question regarding the lawyer’s honesty, trustworthiness, or fitness as a lawyer in other respects shall inform the appropriate authority.

(C) A judge who receives information indicating a substantial likelihood that another judge has committed a violation of this Code shall take appropriate action.

(D) A judge who receives information indicating a substantial likelihood that a lawyer has committed a violation of the Rules of Professional Conduct shall take appropriate action.

Considerations: As the generative AI landscape continues to rapidly evolve, how will a judge know when another judge or lawyer is violating the Colorado Rules of Professional Conduct or the Colorado Code of Judicial Conduct through the improper use of generative AI?

Colorado UPL Rules

The Colorado UPL Rules raise the question of whether a generative AI program can “exercis[e] legal judgment.” These rules define the “[e]xercise of legal judgment” as “the application of actual or purported knowledge or understanding of the law, beyond that of the ordinary citizen, to a particular set of facts.”13

Jurisdiction and Prohibited Colorado UPL Activities

CRCP 232.2 provides:

(c) Prohibited Activities. The unauthorized practice of law by a nonlawyer includes the following:

(1) Exercising legal judgment to advise another person about the legal effect of a proposed action or decision;

(2) Exercising legal judgment to advise another person about legal remedies or possible courses of legal action available to that person;

(3) Exercising legal judgment to select a legal document for another person or to prepare a legal document for another person, other than solely as a typist or scrivener;

(4) Exercising legal judgment to represent or advocate for another person in a negotiation, settlement conference, mediation, or alternative dispute resolution proceeding;

(5) Exercising legal judgment to represent or advocate for another person in a hearing, trial, or other legal proceeding before a tribunal;

(6) Advertising or holding oneself out, either directly or impliedly, as an attorney, a lawyer, “Esquire,” a legal consultant, or a legal advocate, or in any other manner that conveys capability or authorization to provide unsupervised services involving the exercise of legal judgment;

(7) Owning or controlling a for-profit entity that is not authorized under C.R.C.P. 265 and that provides services involving the exercise of legal judgment;

(8) Soliciting any fees for services involving the exercise of legal judgment;

(9) Owning or controlling a website, application, software, bot, or other technology that interactively offers or provides services involving the exercise of legal judgment; and

(10) Performing any other activity that constitutes the practice of law as set forth in subsection (b) above.

Considerations: If a self-represented litigant uses generative AI to draft pleadings or a legal document, such as a separation agreement in a domestic relations case, is the generative AI platform “exercising legal judgment”? Lest this seems far-fetched, ChatGPT has reportedly already generated separation agreements when prompted to do so.14 Notably, at times it declined to do so, responding (correctly) that “as an AI language model, I cannot create legal documents or provide legal advice.”15

This question may increasingly arise as entrepreneurs look for ways to use technology to increase access to justice for self-represented litigants in new ways. For instance, in Florida Bar v. TIKD Services LLC,16 the Florida Supreme Court enjoined respondents—who operated a website and mobile application through which drivers could receive assistance in resolving traffic tickets—from doing business. The court concluded the respondents were in the business of selling legal services to the public and thus engaged in the unauthorized practice of law.17

An AI startup called DoNotPay took things in yet a different direction.18 It intended to have an AI-powered bot provide real-time assistance to a defendant in a traffic case in California in February 2023.19 The startup planned to have the defendant wear smart glasses that would record the court proceedings and through which the bot would “dictate responses into the defendant’s ear from a small speaker.”20 The system purportedly relied on text generators, ChatGPT, and DaVinci (an image-creating AI platform).21 The company abandoned plans to move forward after multiple bar organizations allegedly threatened the company under their rules prohibiting the unauthorized practice of law.22

These businesses raise a question about generative AI and the unauthorized practice of law: Would a tech-savvy entrepreneur violate the Colorado UPL Rules by selling to self-represented litigants in Colorado a generative AI service that creates legal documents the self-represented litigants can file in court? What if the entrepreneur “feeds” legal advice in real time to a self-represented litigant during court proceedings?23 Would the consumer violate the Colorado UPL Rules by using the generative AI platform?

Additionally, does CRCP 232.2(c)(8), which bars a nonlawyer from “[s]oliciting any fees for services involving the exercise of legal judgment,” apply when an AI platform charges the fee for those services?

And, importantly, how do we balance the promise that generative AI holds to increase access to justice with concerns about the unauthorized practice of law, which is prohibited to protect both the public and the integrity of the legal system from unqualified individuals “who provide incompetent legal services”?24

Conclusion

Colorado appears to be one of the first states—if not the first state—to consider whether its existing rules governing professional and judicial conduct and the unauthorized practice of law should be amended given the rise of new, powerful generative AI tools. As noted, these tools hold great potential to help lawyers, clients, judges, and self-represented litigants alike. Generative AI platforms can, among other things, streamline document drafting, save clients money, and increase access to justice. But these tools also present risks to users of this technology and to our system of justice. For instance, court filings produced using generative AI may unwittingly include biased information or hallucinated citations, and a lawyer may unintentionally disclose confidential client information to third parties by including it in a query on certain platforms.

This is why it is important to consider whether amendments to our existing rules are necessary to realize the promises and meet the challenges presented by this rapidly evolving technological landscape. In doing so, we must carefully weigh the need for adaptability and innovation against the important principles that are the foundation for these rules.25

The authors appreciate and acknowledge the invaluable assistance of Abigail Armstrong, Emma Berry, Ellen Blatt, Cormac Bloomfield, Caroline Brown, Madeline Finlayson, Rachel Harrington, Felix Hollaway, Madeline Leibin, Harlan Norby, and Devin Schultze.

---

Explore This Issue

Liberatory Love

President’s Message

By Tiera Brown (guest author)

Gary Abrams, Larry Litvak, Gilbert Martinez, William Nowland, and Raymond Ruth

In Memoriam

Is Modern Standing Doctrine Just Calvinball?

As I See It

By John K. Crisham

Related Topics

Notes