LawMay Issues Its Generative AI Use Policy

The LawMay P.C. Generative AI Use Policy took effect on June 1, 2026 and applies to every attorney and staff member of the firm. Each person signs an acknowledgment, to be re-signed annually.

Why We Wrote It

The immediate reason is risk.

In 2023, in Mata v. Avianca, the Southern District of New York sanctioned counsel and their firm for submitting AI-fabricated cases and quotations to the court. In 2024, in Park v. Kim, the Second Circuit referred an attorney who cited nonexistent authority to its grievance panel.

It has also happened in matters we handle.

In a copyright case we defended in the Southern District of Florida, opposing counsel's filing was identified by the court as containing AI-fabricated citations. That confirmed for us that individual caution is not enough. A firm needs a single, shared process.

If the only goal were to prohibit risk, listing a few bans would be easy. But AI has already entered legal research, summarization, and first drafting. What actually needs answering is: what work may use AI, which tools may be used, who verifies the output, and how problems are handled when they arise.

Those are the questions this policy answers.

What We Considered in Drafting It

Before drafting, we reviewed sixteen sets of rules, decisions, and industry materials — including ABA Formal Opinion 512, the ethics opinions of the California, New York City, and Florida bars, USPTO practice guidance, the relevant federal decisions, and policy templates and judicial guidance from the United States and Singapore. The full list appears at the end.

Drawing the lines took the longest.

On client confidentiality and citation verification we adopted stricter rules. Pre-filing invention disclosures, unpublished patent applications, and trade secrets may not be entered into any external AI tool — not even after anonymization. Cases, statutes, and quotations supplied by AI may not go into a filing unless the attorney has retrieved the original from an authoritative database and verified it.

The remainder we wrote to fit how we actually work.

Many available templates are designed around large-firm staffing: many sections, and a dedicated compliance function to run them. We did not copy that. Filings with three or fewer citations do not require a separate verification checklist; only filings with more than ten require a second attorney's sign-off. Free AI tools are not banned outright, but they may not touch client information.

The resulting main policy runs two pages, with four annexes:

1. a citation verification checklist;
2. an AI drafting workflow;
3. a personnel acknowledgment; and
4. an AI clause for engagement letters.

What the Policy Provides

1. Client information may go only to tools that meet our requirements

Client information may be processed only using commercial paid AI tools the firm has approved.

Before approval, the firm reviews the vendor's terms of service, privacy policy, and data processing agreement, confirming that inputs will not be used to train models, and checking the arrangements for data retention and deletion, access rights, subprocessing, cross-border transfer, and security incident notification.

Tools that have not passed that review may not process client information. Free and personal subscription versions may not touch client information at all.

2. Client information is classified in three tiers

Public information may be processed with AI.

Ordinary client confidential information may be entered only into firm-approved tools.

The most sensitive materials — pre-filing invention disclosures, unpublished patent applications, trade secrets — may not be entered into any external AI tool.

3. Attorneys must verify citations personally

Cases, statutes, and quotations supplied by AI must be retrieved by the attorney from an authoritative database, confirmed to exist, confirmed accurate as quoted, and checked for continued good law.

The verification obligation rests with the signing attorney. Using AI does not transfer it.

4. Clients decide whether AI is used

The firm describes its use of AI in the engagement letter.

A client may ask us to limit AI use to particular stages, or to use no generative AI at all in its matter. The responsible attorney implements that instruction.

5. Problems are reported immediately

On discovering a suspected fabricated citation, client information entered into an unapproved tool, or a possible data breach, staff must report to a partner immediately, without concealment or delay.

The policy is reviewed annually. Attorneys and staff complete annual training and re-sign the acknowledgment.

Finally

AI can participate in research, summarization, and first drafting. What gets filed, and the judgment behind it, remains the attorney's to verify item by item and to answer for.

Clients with specific requirements about AI use in their own matters can raise them directly with the responsible attorney.

Firms of comparable size working on their own policies are welcome to compare notes on what implementation actually involves.

Principal References

Professional rules and official guidance

• ABA Model Rules of Professional Conduct, particularly Rules 1.1, 1.4, 1.6, 5.1, and 5.3
• ABA Formal Opinion 512, Generative Artificial Intelligence Tools (July 29, 2024)
• The State Bar of California, Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law (Nov. 16, 2023); California Rules of Professional Conduct
• New York City Bar Association, Formal Opinion 2024-5, Generative AI in the Practice of Law (2024)
• The Florida Bar, Ethics Opinion 24-1 (Jan. 19, 2024)
• USPTO, Guidance on Use of Artificial Intelligence-Based Tools in Practice Before the United States Patent and Trademark Office, 89 Fed. Reg. 25609 (Apr. 11, 2024)
• Fed. R. Civ. P. 11(b); 37 C.F.R. § 11.18

Decisions

• Mata v. Avianca, Inc., 678 F. Supp. 3d 443 (S.D.N.Y. 2023)
• Park v. Kim, 91 F.4th 610 (2d Cir. 2024)

Industry and comparative materials

• Ministry of Law (Singapore), Guide for Using Generative AI in the Legal Sector (Mar. 6, 2026)
• Singapore Courts, Guide on the Use of Generative Artificial Intelligence Tools by Court Users (effective Oct. 1, 2024)
• ABA Law Practice Division, A Practical Checklist for Using AI Responsibly in Your Law Firm (2026)
• Edward J. Jarot Jr. & William D. Lohrman, AI in the Small Law Firm — The Need for a Formal Written Policy, ISBA The Prompt, Vol. 1 No. 2 (Mar. 2025)
• Catherine Reach, Beyond the Ban: Why Your Law Firm Needs a Realistic AI Policy in 2026, North Carolina Bar Association (Jan. 13, 2026)
• Clio, Law Firm AI Policy Template, Tips & Examples
• Bloomberg Law, Federal Court Judicial Standing Orders on Artificial Intelligence

This article is a general discussion and does not constitute legal advice on any specific matter. Public case information is as reflected in the court records.