Hey Alexa – When Can I Draft Documents With AI?
Kate Gould, Esq.
You might consult Alexa for the daily forecast or the perfect dinner party playlist but leave the ethics of using AI to the ABA. Although AI developments seem to be happening at a frenetic pace and could improve efficiency in your practice, ChatGPT shouldn’t be your sole resource for legal authority or forms.
As lawyers, we spend a significant amount of time drafting documents, including client correspondence, pleadings, and memos. While using AI to prepare these documents may save time, its use can result in the unauthorized disclosure of confidential client information. This article will address some of the ethical considerations related to a lawyer’s decision to use generative AI to assist with everyday tasks. No need to ask your Echo.
The ABA Model Rules
Multiple Rules of Professional Conduct come into play when using AI in your legal practice. As detailed below, consider how these Model Rules may be implicated:
Competency (Rule 1.1)
ABA Model Rule 1.11 requires that a lawyer provide competent representation to a client. What does this mean? You must possess the legal knowledge, skill, thoroughness, and preparation reasonably necessary for the representation. If you are going to use generative AI in your practice, you must understand how it works and what its limitations are to avoid any potential harm to your client’s interests. And as further discussed below, you must verify the authenticity of any purportedly valid legal information or case citations generated by an AI model like ChatGPT. False or incomplete information, known as “hallucinations,” can be produced by AI and appear to be legitimate.
Client Communications (Rule 1.4)
This Rule requires a lawyer to “reasonably consult with the client about the means by which the client’s objectives are to be accomplished.” If you intend to use AI as one means to accomplish your client’s objectives, you should “promptly inform” the client of that intention in order to obtain the client’s informed consent. According to ABA Resolution 1122, a lawyer should obtain approval from the client before using AI, and this consent must be informed. You might use your initial fee agreement to ensure you have client consent from the outset of the case. Interestingly, the Resolution also states that if you decide not to use AI, you must communicate that to the client if using it would benefit them.
Confidentiality (Rule 1.6)
ABA Model Rule 1.6(a)3 provides that “[A] lawyer shall not reveal information relating to the representation of a client unless the client gives informed consent, the disclosure is impliedly authorized in order to carry out the representation, or the disclosure is permitted by paragraph (b).” As such, disclosing any confidential client information to an AI tool without the client’s informed consent would constitute a violation of the Rule. Even with their consent, the use of an outside AI model is problematic. For example, if you used it to draft an email in reply to opposing counsel, you would be disclosing your strategy by asking the model to craft a response based on your legal position – even if you somehow avoided plugging in confidential client information. Quite frankly, the risks of using outside generative AI when it comes to the question of confidentiality may be the biggest hurdle to its use.
Candor to the Tribunal (Rule 3.3)
You’ve probably heard (or maybe Siri mentioned it) that lawyers are routinely being reprimanded for using ChatGPT or another AI model for legal research and then citing fake cases in their briefs. Model Rule 3.3(a)(1)4 prohibits a lawyer from knowingly making “a false statement of fact or law to a tribunal or fail[ing] to correct a false statement of material fact or law previously made to the tribunal by the lawyer.” Lawyers simply cannot rely on AI models alone for the case law they need given the risk of hallucinations. Instead, as with any case you might include in a brief, you must Shepardize the case to ensure it is still valid precedent – not to mention a legitimate case. So, whether you find your case “in the books,” on Westlaw or Lexis, or via an AI generative model, fulfill your ethical responsibility to confirm the validity of the case law you present to the court.
When Can You Use Generative AI to Draft Documents?
If you know you can comply with the above Rules – and your client has given you the requisite informed consent – there are some circumstances were you can use generative AI to draft documents.
- The first draft of a litigation hold letter – As long as you do not input any client or matter information and tailor the letter to your needs, you could use AI for the initial draft of your letter.
- The first draft of a contract – Sometimes, working from a form is easier than starting from scratch. After ChatGPT generates a draft contract, you can add the necessary client information and revise it to fit your needs and comply with the laws of your state.
- Initial template creation – Whether you are developing an internal memo form or client intake questionnaire, you could ask AI to get you started.
- Initial drafts of discovery requests – Preparing discovery requests can be tedious work. Having AI produce an initial draft will leave more time for you to develop case-specific Interrogatories and Requests for Production. Be sure to check your local rules to ensure you comply with all form requirements and do not exceed to maximum number of questions or requests.
- Marketing materials – It is difficult to find the time to generate timely blog posts or client newsletters while managing a heavy caseload. If you use an AI model to produce these materials, be sure to check the case citations!
The potential benefits of using AI will likely eventually outweigh the risks as it continues to develop. Further, as it becomes more commonplace to have captive AI tools where client confidentiality can be sufficiently guaranteed within the firm, we will likely see the use of AI further grow into our day-to-day practices. However, lawyers will always have the responsibility to protect their client’s interests, whether our work product is aided by AI-generated content or not. And, now that you know the ethical considerations to incorporating AI into your practice, your virtual assistant can get back to setting alarms and finding recipes.
Kate is a Risk Management Consultant who enjoys providing risk management resources to Attorney Protective’s insured attorneys. She is looking forward to summer in the Midwest (when she isn’t looking up file retention guidelines around the country).
Get Lawyers Liability Insurance Malpractice Insurance Quote
This document should not be construed as legal advice. Because the facts applicable to your situation may vary, or the laws applicable in your jurisdiction may differ, please contact your attorney or other professional advisors if you have any questions related to your legal or medical obligations or rights, state or federal laws, contract interpretation, or other legal questions. Attorney Protective is a MedPro Group/Berkshire Hathaway company that protects the reputations and assets of attorneys across the nation. “Attorney Protective” and “AttPro” are also used as marketing names used to refer to the attorney professional liability programs administered by Attorney Protective or its affiliates and underwritten by National Liability & Fire Insurance Company, AttPro RRG Reciprocal Risk Retention Group, National Fire & Marine Insurance Company – each of which has earned an A++ financial strength rating from AM Best – or other Berkshire Hathaway affiliates. Product availability is based upon business and/or regulatory approval and may differ among companies. © Attorney Protective. All Rights Reserved. AttPro-240626