Client Communication Protocols Regarding AI Use

Transparent and effective communication with clients about the use of AI in their matters is essential for maintaining trust, managing expectations, and fulfilling ethical obligations, particularly the duty to communicate under TDRPC Rule 1.03.

Key Elements of Client Communication Protocols for AI

Timing of Disclosure

• Ideally, initiate the discussion about potential AI use early in the representation, perhaps during the initial consultation or within the engagement letter.
• Revisit the discussion as needed if the scope of AI use changes significantly or new AI tools are introduced that could impact the client's matter.

What to Disclose

• Intended Use. Clearly explain how AI tools might be used in the client's specific matter (e.g., "We may use AI software to assist with reviewing the large volume of documents in your case," or "AI may help draft initial responses to standard discovery requests, which will then be thoroughly reviewed by an attorney").
• Benefits. Articulate the potential advantages for the client, such as increased efficiency, faster turnaround times for certain tasks, or potentially reduced costs.
• Risks and Limitations. Be transparent about the downsides:
  • Confidentiality. Explain measures taken to protect their data when using AI, especially with third-party vendors. If any risk exists despite precautions, inform the client.
  • Accuracy. Explain that AI can make mistakes ("hallucinate") and that outputs are not guaranteed to be accurate or complete without human review. Emphasize the firm's verification process. Mention potential knowledge cut-offs of large language models if relevant to research tasks.
  • Bias. Briefly mention the possibility of AI bias and the firm's efforts to mitigate it.
  • Safeguards. Describe the steps the firm takes to mitigate risks, such as the mandatory attorney review process, vendor vetting, and data security measures.

Obtaining Informed Consent

• For uses of AI that constitute a "significant means" of representation or involve sharing client data with third parties, obtaining the client's informed consent is crucial, both ethically and for risk management.
• While verbal consent is possible, documenting consent in writing (e.g., in the engagement letter, a separate addendum, or an email confirming a conversation) is strongly recommended.
• Ensure the client understands the information provided and has the opportunity to ask questions.

Format and Tone

• Use clear, accessible language. Avoid technical jargon.
• Tailor the explanation to the individual client's understanding and the complexity of their matter.
• Maintain a professional and reassuring tone, emphasizing that AI is a tool used under the attorney's control and supervision.

Handling Client Objections or Concerns

• Be prepared to discuss alternatives or agree not to use AI for specific tasks if the client expresses discomfort or has valid privacy concerns. The attorney's duty is to represent the client diligently and according to their objectives (within legal and ethical bounds).

Documenting Communications

• Keep records of when and how AI use was discussed with the client and any consent obtained. This documentation is vital for demonstrating compliance with ethical duties.

Updates

• If the firm's AI practices change in a way that significantly impacts the client's matter or data (e.g., switching to a new AI vendor, using AI for a different type of task), provide an update to the client.

• Develop Standard Disclosures. Draft template paragraphs for engagement letters or standard email language to initiate the conversation about AI use.
• Train Attorneys and Staff. Ensure everyone who interacts with clients knows how to discuss AI use transparently and answer common questions.
• Identify "Significant Uses." Establish internal guidelines on which AI uses are significant enough to require explicit client discussion and consent.
• Obtain Written Consent. Implement procedures for obtaining and documenting written consent whenever feasible for significant AI uses.
• Be Prepared for Questions. Anticipate client concerns about privacy, accuracy, and cost.
• Document Discussion. Record details of AI discussions in the client file.