A firm without a formal AI Acceptable Use Policy is operating without one of the basic instruments a profession of this risk profile requires. Shadow AI use, where attorneys and paralegals adopt consumer-grade tools on personal devices to save time, is already happening in most firms that have not put a policy in place. The exposure is real and largely unmonitored.

A working AI policy does not slow innovation. It clears the path for it by telling people exactly which tools are sanctioned, which are not, and what the verification expectations are.

What an AUP Has to Cover

State bar task forces and ethics guidance have converged on a common set of elements an AUP must address.1, 2 The policy begins with clear definitions, separating basic software from generative AI, which the policy should define as computer-based systems that produce new content from underlying models trained on large datasets.

State agencies are publishing model frameworks worth studying. The Louisiana Division of Administration has issued an Artificial Intelligence Acceptable Use Policy3 that gives both government and professional organizations a usable starting reference.

A functional AUP answers three core questions for every person in the firm:

  1. Which tools are approved? The policy should name the sanctioned closed systems for client work and explicitly prohibit public, open-source tools for anything involving client information.
  2. Which tasks are permitted? The policy should clarify that AI may be used to draft, summarize, and outline, and should prohibit any use of AI to manipulate evidence, fabricate content, or generate work that misrepresents authorship to a tribunal.
  3. What is the human verification standard? Under Rules 3.1 and 3.3, attorneys have a duty to verify the accuracy of AI-generated material. The policy must make verification mandatory before any AI-assisted output leaves the firm.

Supervision Under Rules 5.1 and 5.3

The policy must also address supervisory responsibility. Rules 5.1 and 5.3 hold partners and supervisory lawyers accountable for the conduct of their subordinate attorneys and non-lawyer staff. A paralegal who uses an unapproved AI tool to fabricate a citation puts both herself and the partner who signed the brief into disciplinary territory. The policy is the vehicle that makes the supervisory expectation explicit, documented, and enforceable.

Sources

  1. New Jersey State Bar Association. Task Force on Artificial Intelligence (AI) and the Law / Preliminary Guidelines on New Jersey Lawyers' Use of Artificial Intelligence. Guidelines for ethical AI use by attorneys, covering AUP requirements, supervision, and disclosure.
  2. New Jersey State Bar Association. AI Task Force Guidelines (supplementary ethics opinions).
  3. Louisiana Division of Administration. Artificial Intelligence Acceptable Use Policy. Official state government AI policy framework addressing authorized tools, prohibited uses, and data security requirements.

Ready to bring responsible AI to your firm? Let's start with a conversation.

Book a Discovery Call