Technology is interwoven with the practice of law, transforming the way lawyers work and interact with clients. One of the most significant advancements in technology is generative artificial intelligence (GenAI), which has the potential to revolutionize many aspects of legal practice. However, the adoption of GenAI also brings with it a host of ethical and professional considerations.

It has been almost two years since the public introduction of ChatGPT, the large language model (LLM) created by OpenAI. ChatGPT was adopted at a rate never before seen, reaching 100 million active monthly users in only two months.¹ By comparison, TikTok and Instagram took nine months and 2.5 years respectively to hit that same milestone.² GenAI technology has been implemented into the legal industry at an unprecedented pace. The power of the technology and the speed of adoption have fostered lawyers’ and regulators’ curiosity and apprehension. The number of ethics opinions and guidance published over the past year is evidence of the unprecedented effect GenAI has had on the legal industry already and the anticipation of what is to come.³

Generative Artificial Intelligence (GenAI)

What is GenAI? AI does not have a set definition. It is usually understood to mean computer processes, tasks, and outputs that are thought to require some level of human intelligence to complete. GenAI is a specific type of AI in which the model can generate new predictive output – text, images, audio, video – based on patterns and algorithms in the data it has been trained on. The ability for users to input instructions or requests, called prompts, in ordinary language and interact with the GenAI users in a conversational format make it very easy to use and is the basis, in part, for the unprecedented rate of adoption by the public and the legal community.

The terms GenAI and LLMs are often used interchangeably. The major distinction is that LLMs are a subset of GenAI. LLMs focus on the generation of text output, whereas GenAI output can be text, images, video, audio, or programming code.

GenAI models are trained on vast datasets and use advanced machine learning techniques to generate text, answer questions, and perform various tasks, all in natural language. These capabilities make GenAI a powerful tool for legal professionals, enabling them to streamline processes and improve efficiency. However, rapid adoption without understanding the strengths and limitations of the tool has led to some misuse of the tool in the practice of law.

Cautionary Tales

Most people have now heard stories about the numerous instances in which lawyers used public GenAI models to prepare filings with a court without verifying the generated output, leading to significant misrepresentations to the court. These cases highlight the importance of checking the accuracy of GenAI outputs because failing to do so violates the duty of candor to the tribunal.

For example, there have been cases in which lawyers submitted court documents containing fabricated case citations generated by GenAI. Such mistakes not only undermine the lawyer’s credibility but also have serious legal consequences. However, it must be stressed that GenAI is one kind of tool that when understood and used correctly can enhance – not replace – the efficiency and quality of the legal services delivered. The tool has its weaknesses, and it is the lawyer’s duty to understand the benefits and the risks of this technology. Therefore, this article readdresses the strengths and weaknesses of GenAI and the importance of verifying GenAI outputs, which is the lawyer’s duty no matter where the output came from.

Ethical Guidance from Bar Associations

These cautionary tales put the spotlight on the risks of lawyers using GenAI in the practice of law. Lawyers have an obligation to understand and abide by the ethical rules under Wisconsin’s Rules of Professional Conduct, codified in Supreme Court Rules (SCR) chapter 20. The importance of adhering to ethical rules when using GenAI has now been addressed by various state bar association ethical opinions⁴ or guidance.⁵ Most recently, the American Bar Association’s Standing Committee on Ethics and Professional Responsibility issued Formal Opinion 512 on the use of GenAI.⁶

The following are specific provisions of SCR chapter 20⁷ that must be considered when determining whether GenAI should be integrated into a law practice. The common themes across the opinions from bar associations are that lawyers must understand AI and stay educated about the benefits, risks, and limitations (SCR 20:1.1 Competence); take reasonable steps to prevent unauthorized disclosure (SCR 20:1.6 Confidentiality); supervise GenAI use by attorneys, staff, and third-party vendors (SCR 20:5.1 and 20:5.3 Supervision of Subordinates); verify GenAI outputs before submitting to courts (SCR 20:3.1, 20:3.3, and 20:8.4(c) Meritorious Claims and Candor to the Court); communicate the use of GenAI to clients when necessary or asked (SCR 20:1.4 Communication with Clients); inform clients about AI-related billing and the duty to charge reasonable fees (SCR 20:1.5 Fees); and be aware of the potential unauthorized-practice-of-law issues (SCR 20:5.5).

Bar associations generally are taking the approach that the current ethical rules are sufficient in addressing the potential risks involved with the use of GenAI. So, given what has been learned in 2023 and 2024 and the current guidance regarding GenAI, what practical steps should lawyers take when implementing GenAI into a law practice?

Considerations and Best Practices

Competency is a fundamental requirement for lawyers. According to SCR 20:1.1, lawyers must have the legal knowledge, skill, thoroughness, and preparation reasonably necessary for the representation of their clients. SCR 20:1.1, Comment 8, clarifies that competence includes “understanding the benefits and risks associated with relevant technology.” This means that lawyers must make reasonable efforts to stay updated and educated on technological advancements, such as GenAI, that could affect their practice.

Education on the benefits and risks of technology in the law practice is key to lawyer compliance with SCR 20:1.1. With GenAI it is exceedingly difficult to stay on top of the fast-paced changes taking place. Lawyers should become educated about the benefits and risks of GenAI by obtaining continuing legal education about GenAI and ethics, reading articles, listening to podcasts, and using GenAI.

Benefits of GenAI

As with any tool, GenAI has its strengths and limitations. It is important to understand how GenAI can enhance legal practice so these benefits can be weighed against any potential risks. Some of the many areas in which GenAI can be used, and related benefits, are listed below.

• Legal Research. Although outputs must be verified, the speed and accuracy of the research task can be improved by GenAI. Publicly available GenAI (for example, ChatGPT, Gemini, and Anthropic) should not be considered legal research tools. If lawyers use these public models, all outputs must be scrutinized and verified. A safer tool would be one of the GenAI models integrated into legal-research-specific products (for example, Lexis + AI, Westlaw Precision, and vLex VincentAI). These tools limit the universe of information to their vast and trusted databases of legal authority and therefore reduce the frequency of hallucinations. Verification of all outputs is necessary even with the legal-specific models because nuances in the law are prevalent and therefore inaccurate results are possible.

• Drafting or Reviewing Documents. GenAI can create initial drafts of legal documents that lawyers can review and customize. Some products can analyze a user’s document library and create legal documents in the style and format consistent with past documents. Again, although offerings from legal-specific GenAI are recommended (for example, Spellbook, Casetext CoCounsel), Microsoft Copilot for Microsoft 365 is also able to perform drafting tasks and can be limited to a user’s data within the user’s system. Practice management software providers are also starting to implement GenAI models into their systems, allowing users to limit the GenAI to the database of their own law firm documents and data already stored with the practice management software.

• Client Communication. GenAI can help draft general client communications and updates. It can assist with clarity and tone to improve communications with clients.

• Marketing and Education. GenAI can generate content for newsletters, blogs, social media posts, and client educational materials. For example, GenAI can assist with summarizing and distilling the process that can be expected in a client matter into an informational one-page document presented at an understandable level for the client.

• Automating Tasks. GenAI can assist with automating routine tasks, such as scheduling, billing, and the client intake process.

• Firm Policy and Procedure Generation. GenAI can assist with creating first drafts of law firm policies. Law firms can use GenAI as a starting point for generating information security policies, GenAI use policies, and employee policies and procedures.

• Idea Machine. GenAI can be used to help generate ideas to combat writer’s block or to evaluate whether there is anything that a user has not thought about when writing about a topic.

• Increased Efficiency. Adopting some or all of these suggestions for how GenAI could benefit a law practice allows lawyers to free up more time that can be used to provide better client service. Using the technology as a tool to assist with time-consuming back-end tasks enables the lawyer to provide better in-person communication with clients, leading to greater client satisfaction.

Risks of GenAI

Despite its benefits, GenAI poses several risks and has limitations that must be considered.

• Confidentiality. There is a risk of inadvertent disclosure of confidential information if lawyers are uploading it to GenAI models. The LLMs backing the GenAI products learn on data. Many of the public GenAI products use inputted data from users to continue the training of the models.

• Hallucinations. GenAI can generate inaccurate or completely made-up information.

• Bias. There is potential for bias in the training data used to develop the GenAI model, leading to biased outputs.

• Improper Prompting. Poorly crafted prompts can result in poor outputs.

• Vendor Terms. Different GenAI vendors have different terms of service and data-handling policies, which lawyers must understand to protect client information.

• Unauthorized Practice of Law. As noted by the Minnesota State Bar Association Working Group on AI, historically only humans were capable of applying law to specific facts but today’s LLMs are now capable of making similar analysis, raising concerns over GenAI being relied on for legal advice.⁸ Although these LLMs appear capable of legal analysis supplanting professional judgment for the output of LLMs (which is based on predictive patterns rather than true reasoning and legal analysis), serious concerns are raised regarding the unauthorized practice of law.

• Supervision of GenAI Use. Supervising lawyers must ensure that all GenAI users in a law practice are trained on and comply with GenAI policies and ethical requirements.

• Copyright Concerns. Clients with copyright concerns might not want their information inputted into GenAI out of concerns about losing copyright protection. There are numerous ongoing court cases regarding GenAI and copyrighted materials.⁹

To mitigate these risks, lawyers must be diligent in their use of GenAI, ensuring that they understand the technology and implement appropriate safeguards. This includes regular training, robust policies, and continuous monitoring of GenAI outputs.

Best Practices for Getting Started with Implementing Generative AI into a Law Practice

To maximize the benefits of GenAI while mitigating its risks, lawyers must adopt a proactive and informed approach. This includes regular training, continuous monitoring, and adherence to ethical guidelines.

The following are some recommended best practices when considering whether to use GenAI and then implementing it into a law practice:

• Keep current with education and knowledge about the strengths and limitations of GenAI.

• Become familiar with GenAI by purchasing a paid, public version of ChatGPT and using it outside the law practice to generate ideas, plan vacations, lists, and generate suggestions in everyday life.

• Understand the terms of service of the GenAI model.

• Avoid uploading confidential information.

• Use legal-specific GenAI tools that are designed with good-quality protections and controls.

• Disclose the use of GenAI to clients and courts when necessary and required.¹⁰

• Recognize that GenAI may be reasonable to use with certain clients or certain matters but that the reasonableness analysis must be done for each client and each matter.

• Create and enforce a GenAI use policy.

• Educate all law firm users and supervise attorneys and staff in the use of GenAI within the firm.

• Understand how to effectively create prompts to return better and more accurate output from GenAI.

• Be aware of the potential for bias in GenAI outputs.

• Understand that billing for time saved when using GenAI will result in unreasonable fees.

• Discuss how the law firm will address with clients the firm’s decision to use or not use GenAI.

• Always verify all GenAI outputs before using them in court filings or other legal documents.

Implementing these practices ensures that lawyers use GenAI responsibly and ethically, maintaining the trust of their clients and upholding professional standards. Choosing the right tool depends on the specific needs of the legal practice and the nature of the tasks performed. By selecting tools designed for legal use, lawyers can benefit from enhanced protections and better alignment with ethical requirements.

Conclusion

GenAI continues to present both opportunities and challenges for the legal profession. By understanding and adhering to ethical rules and guidance, lawyers can harness the power of GenAI to enhance their practice while safeguarding client confidentiality and maintaining professional integrity. As technology continues to evolve, ongoing education and vigilance are essential to navigate this dynamic legal landscape effectively.

Readers who have questions about implementing GenAI into their law practices can reach out to Practice411™, the State Bar of Wisconsin’s Practice Management Program, by visiting wisbar.org/practice411 or emailing practicehelp@wisbar.org.

Endnotes

1 Benj Edwards, ChatGPT Sets Record for Fastest-Growing User Base in History, Report Says, Ars Technica (Feb. 1, 2023), https://arstechnica.com/information-technology/2023/02/chatgpt-sets-record-for-fastest-growing-user-base-in-history-report-says/.

2 Id.

3 These recommendations were appropriate when this article was written but given the speed at which things are changing regarding GenAI and the legal industry, some of the information might change or become obsolete. Readers should check current recommendations for using GenAI in law practices.

4 N.J. Cts., Supreme Ct. of N.J., Preliminary Guidelines on New Jersey Lawyers’ Use of Artificial Intelligence,https://www.njcourts.gov/sites/default/files/notices/2024/01/n240125a.pdf; Pa. Bar Ass’n Comm. on Legal Ethics & Pro. Resp. & Philadelphia Bar Ass’n Pro. Guidance Comm. Joint Formal Op. 2024-200, Ethical Issues Regarding the Use of Artificial Intelligence, https://www.pabar.org/Members/catalogs/Ethics Opinions/Formal/Joint Formal Opinion 2024-200.pdf; State Bar of Mich., Ethics JI -155, https://www.michbar.org/opinions/ethics/numbered_opinions/JI-155; Fla. Bar Ethics Op. 24-1, https://www.lawnext.com/wp-content/uploads/2024/01/FL-Bar-Ethics-Op-24-1.pdf; State Bar of Cal. Standing Comm. on Pro. Resp. & Conduct, Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law, https://www.calbar.ca.gov/Portals/0/documents/ethics/Generative-AI-Practical-Guidance.pdf.

5 Minn. State Bar Ass’n, Working Grp. on AI, Implications of Large Language Models (LLMs) on the Unauthorized Practice of Law (UPL) and Access to Justice, https://www.mnbar.org/docs/default-source/default-document-library/msba-ai-working-group-final-report-and-recommendations.pdf; State Bar of Tex., Taskforce for Responsible AI in the Law, Interim Report to the State Bar of Texas Board of Directors, https://www.texasbar.com/AM/Template.cfm?Section=Meeting_Agendas_and_Minutes&Template=/CM/ContentDisplay.cfm&ContentID=62597; N.Y. State Bar Ass’n Task Force on Artificial Intelligence, Report and Recommendations to NYSBA House of Delegates, https://nysba.org/app/uploads/2022/03/2024-April-Report-and-Recommendations-of-the-Task-Force-on-Artificial-Intelligence.pdf.

6 ABA Standing Comm. on Ethics & Pro. Resp., Formal Op. 512, Generative Artificial Intelligence Tools (July 29, 2024), https://www.americanbar.org/content/dam/aba/administrative/professional_responsibility/ethics-opinions/aba-formal-opinion-512.pdf.

7 Wisconsin Supreme Court Rules Chapter 20, https://www.wicourts.gov/sc/scrule/DisplayDocument.pdf?content=pdf&seqNo=749164,https://www.wicourts.gov/sc/scrule/DisplayDocument.pdf?content=pdf&seqNo=675950.

8 Minn. State Bar Ass’n, Working Grp. on AI, Implications of Large Language Models (LLMs) on the Unauthorized Practice of Law (UPL) and Access to Justice (June 2024), https://www.mnbar.org/docs/default-source/default-document-library/msba-ai-working-group-final-report-and-recommendations.pdf.

9 Joe Panettieri, Generative AI Lawsuits Timeline: Legal Cases vs. OpenAI, Microsoft, Anthropic, Nvidia and More, Sustainable Tech Partner (Aug. 5, 2024), https://sustainabletechpartner.com/topics/ai/generative-ai-lawsuit-timeline/.

10 See Ropes & Gray, Artificial Intelligence Court Order Tracker, https://www.ropesgray.com/en/sites/artificial-intelligence-court-order-tracker. (This is an interactive map tool to keep track of the court orders issued in various jurisdictions in the U.S.)