Can Lawyers Use AI to Write Client Emails? What the Bar Rules Actually Say

As AI email tools have become a fixture in legal practice, bar associations have moved from cautious observation to active guidance. The ABA's AI task force has been producing formal opinions, several state bars have issued jurisdiction-specific guidance, and attorneys across the country are asking the same question: am I allowed to use AI to draft the emails I send to clients?

The answer — based on a close reading of the ABA Model Rules and formal opinions issued through mid-2026 — is yes, with conditions. Let's work through each relevant rule carefully.

Rule 1.1 — Competence: Does Using AI Make You Incompetent?

ABA Model Rule 1.1 requires lawyers to provide competent representation, which includes "the legal knowledge, skill, thoroughness and preparation reasonably necessary." Comment 8 specifically addresses technology: lawyers must "keep abreast of changes in the law and its practice, including the benefits and risks associated with relevant technology."

This comment cuts both ways. An attorney who refuses to understand AI tools may eventually be found to lack competence — particularly as AI-assisted drafting becomes standard practice. But an attorney who blindly deploys AI without review is equally at risk.

For email specifically, the competence bar is relatively accessible. Client update emails, billing reminders, and appointment confirmations are lower-stakes than legal memos or court filings. An attorney who reviews an AI draft and confirms its accuracy is satisfying Rule 1.1.

Rule 1.4 — Communication: Can AI Satisfy Your Duty?

Rule 1.4 requires attorneys to keep clients "reasonably informed" and to "promptly comply with reasonable requests for information." It also requires explaining matters "to the extent reasonably necessary to permit the client to make informed decisions."

This rule creates both an argument for and against AI-assisted email. The argument for: AI tools help attorneys communicate more frequently and more promptly, which directly serves the Rule 1.4 duty. Many solo practitioners and small firms under-communicate with clients not from neglect, but because drafting individual status updates for every active matter is genuinely time-consuming.

The argument against careless use: an AI-generated email that contains inaccurate information about a client's case — wrong dates, incorrect procedural status, or inaccurate legal positions — could directly harm the attorney-client relationship and potentially expose the client to prejudice. The duty of communication is about the quality and accuracy of information conveyed, not just the act of sending something.

Rule 5.3 — Supervision: Who Is Responsible for the AI?

Rule 5.3 requires attorneys to supervise the work of nonlawyer assistants and to be responsible for their conduct when it would constitute a violation if done by a lawyer. While the rule was written for human assistants, bar ethics opinions have increasingly applied its principles to AI tools.

The ABA's formal guidance treats AI output the way it treats work product from a supervised paralegal: the attorney is responsible for every word that goes out under their name, regardless of who — or what — generated the first draft.

Privilege Considerations: The Risk Most Attorneys Overlook

Attorney-client privilege is separate from the ethics rules above, but it intersects with AI email tools in ways that deserve specific attention. The core risk: inputting confidential client information into an AI tool that lacks proper data isolation could constitute a disclosure that inadvertently waives privilege.

This isn't hypothetical. Early legal malpractice cases have raised arguments that using a general-purpose cloud AI tool — one whose terms allow data retention or use for model training — constituted a failure to adequately protect confidential communications.

• General-purpose AI tools (free tiers) should never be used for client-specific email drafting. Their data handling terms are not designed for legal confidentiality and may use your inputs for model improvement.
• Enterprise versions of general tools are better, but still lack privilege-specific design features — no privilege scoring, no attorney ethics review on outputs, and no bar-compliance awareness built in.
• Purpose-built legal AI tools with contractual data isolation commitments, signed DPAs, and privilege-aware design are the appropriate choice for matter-specific client communications.

Safe Practices for AI-Assisted Legal Email in 2026

Synthesising the rules above and current bar guidance across jurisdictions, here are the practices that put you on solid ethical ground:

• Use a tool with a signed DPA. Your data processing agreement should explicitly prohibit the vendor from using your inputs for model training and commit to isolated data environments. Get this in writing — not just in a privacy policy.
• Always review before sending. AI drafts are first drafts. You are responsible for every word. This is non-negotiable under Rule 5.3.
• Never rely on AI for substantive legal advice to clients. AI can help you communicate procedural updates, scheduling, and billing. It should not generate substantive legal opinions or strategy recommendations for clients — that is your professional role.
• Disclose if required by your state bar. Several state bars are moving toward disclosure requirements for AI-generated client communications. Check your jurisdiction's latest ethics opinions — California, New York, and Florida have all issued guidance in 2025–2026.
• Maintain audit logs. Keep records of AI-assisted communications in your matter files. This protects you in malpractice disputes and demonstrates the supervisory diligence required under Rule 5.3.
• Choose legal-specific tools over general ones. The privilege, confidentiality, and bar-compliance considerations in legal email are sufficiently unique that general-purpose AI tools create unnecessary risk — even when they perform well on the writing itself.

Bottom Line

Using AI to draft client emails is ethically permissible under the ABA Model Rules, provided you: understand the tool's capabilities and limitations (Rule 1.1), ensure communications are accurate and complete (Rule 1.4), review all output before it goes to clients (Rule 5.3), and use a tool with appropriate data protections to preserve confidentiality and privilege.

The ethical risk isn't in using AI — it's in using it carelessly, or using a tool not designed for the specific requirements of legal practice. The right AI email tool for lawyers is one built around those requirements from the ground up, with privilege scoring, bar ethics review, and data isolation as core features — not afterthoughts.