TAPAs: Using Generative AI to Draft a Brief or Memorandum

The recent evolution of ChatGPT has caused a significant increase in the adoption of artificial intelligence (AI) tools for various tasks within the legal profession, including legal research, document analysis, and drafting. Generative AI, in particular, has gained attention for its ability to generate text that often reads as good as or better than that written by a human. As a result, attorneys have started to use it to draft briefs, memoranda, and other legal documents. While generative AI offers benefits in terms of saving time and effort, it carries with it risks inherent to using generative AI in legal drafting, at least as of the present time.

The risks associated with using generative AI to draft legal memoranda and briefs include:

1. Accuracy and quality concerns. Generative AI models may not always produce accurate or high-quality text, especially in complex legal contexts. AI has a history of generating erroneous statements of law. It has also misinterpreted precedents, which could undermine the credibility of legal arguments. In some cases, AI has actually fabricated authority to support an argument (this tendency of AI is sometimes called “hallucination”).
2. Ethical considerations. Using AI in legal drafting raises ethical concerns regarding transparency and accountability. While the rules and standards vary among jurisdictions, we believe that attorneys should always inform their clients about the use of AI tools in their representation and the limitations of AI-generated content, particularly in terms of legal accuracy and liability.
3. Confidentiality and data security. Generative AI systems typically require access to vast amounts of data to train and operate effectively. Attorneys must consider the potential risks to client confidentiality and data security when using AI tools hosted on external platforms or cloud services, especially if they will deal with sensitive client information in connection with the subject matter.
4. Lack of legal expertise. While AI can assist in drafting text based on existing legal templates and patterns, it currently lacks the nuanced understanding of legal principles and strategies possessed by experienced attorneys. Relying solely on AI-generated content without human review may result in overlooking critical legal issues or failing to tailor arguments to specific case circumstances. That said, the landscape continues to change as a recent iteration of ChatGPT actually passed a bar examination without attending a single day of law school.
5. Regulatory compliance. Attorneys using generative AI in legal drafting must comply with professional conduct rules, including those related to competence, diligence, and client communication. Compliance requirements vary from jurisdiction to jurisdiction, and many jurisdictions have not yet responded to the use of AI. Attorneys must familiarize themselves with the requirements imposed by the jurisdiction in which they practice respecting the use of AI. Failure to exercise proper oversight and judgment in utilizing AI tools and/or failing to comply with applicable regulations could lead to disciplinary action against the attorney and/or legal malpractice claims against the attorney and the law firm.

Below, then, are tips for best practices for attorneys using generative AI.

Tip 1. Understand the Technology

Attorneys must invest time and effort to understand generative AI and how it works. That knowledge must extend to its capabilities, limitations, and potential biases. This will enable attorneys to assess the suitability of AI tools for specific drafting tasks and allow them to make informed decisions about when and how to ethically, effectively, and efficiently use generative AI tools.

Tip 2. Review and Verify Work Product

Attorneys must review and verify all AI-generated content for accuracy, coherence, and relevance prior to incorporating it into legal documents. Attorneys should compare AI-generated text against authoritative legal sources, conduct independent research, and possibly even seek peer review to ensure the quality and integrity of their AI-assisted work product. Under no circumstances should an attorney allow an AI-generated argument to go out of the office prior to the verification of all authority cited, both for legitimacy and appropriate use.

Tip 3. Maintain Control Through Human Oversight

Although AI can streamline the drafting process, ensuring factual accuracy and legal correctness and effectiveness of drafted documents requires oversight and review by an experienced attorney. Attorneys must remain actively engaged in the drafting process, critically evaluating all AI-generated content and revising it as necessary to ensure competent representation for their clients.

Tip 4. Protect Confidentiality and Take Appropriate Steps to Maintain Data Security

Attorneys must prioritize data security and confidentiality when using generative AI tools. This includes selecting reputable AI providers, implementing robust encryption and access controls, and regularly updating security protocols to mitigate the risk of data breaches or unauthorized access.

Tip 5. Educate and Inform Clients

Attorneys should transparently communicate with clients about the use of AI tools in legal drafting, explaining their benefits, limitations, and potential implications. Attorneys should inform and educate their clients about the steps taken to ensure the accuracy of AI-generated content and the steps taken to protect client confidentiality when employing the assistance of AI. As an added precaution, we recommend that you give your clients the opportunity to review and provide input and feedback on drafted documents. Do everything you can to make sure your clients understand and feel that they remain an important part of the process.

Generative AI offers a promising opportunity to enhance the efficiency and productivity of legal drafting. It also presents inherent risks related to accuracy, ethics, confidentiality, expertise, and regulatory compliance. By adopting best practices such as those outlined in this column, attorneys can mitigate the risks and leverage the benefits of generative AI in their legal practice. That said, we will share with you that while the quality of AI work product has continued to improve and it now generates pretty fair writing quality, we have not yet seen a brief or memorandum drafted by AI that we would sign without modification and/or editing.