Arkansas
informalSummary
Arkansas has no standalone formal ethics opinion on AI but has acted through court orders and rule amendments. Administrative Order No. 25 (2025 Ark. 117) governs confidential information in AI tools, and amendments to RPC Rule 5.3 and the Preamble (2025 Ark. 115) place AI outputs squarely within the supervising attorney's responsibility. Two 2025 enforcement actions, In re McClain (Arkansas Supreme Court) and Hatfield v. Pirani (W.D. Ark., $1.5M+ sanctions), have produced concrete consequences.
Applicable ABA Model Rules
- Rule 1.1
- Rule 1.3
- Rule 1.4
- Rule 1.5
- Rule 1.6
- Rule 3.3
- Rule 5.3
Carrier Implications
No Arkansas-specific carrier bulletin identified. The two 2025 enforcement actions signal that carriers offering LPL to Arkansas firms may begin requiring AI use policies, analogous to cybersecurity protocols. Firms with documented AI governance programs are better positioned in both underwriting and claim defense.
Arkansas has moved faster than most smaller states on AI governance through court orders and rule amendments rather than a standalone ethics opinion. Administrative Order No. 25 prohibits disclosure of confidential or sealed court information to external generative AI tools, and the 2025 amendment to Rule 5.3 (re-titled “Nonlawyer Assistance”) makes the supervising attorney personally responsible for AI-generated work product. The amended Preamble forecloses any “the computer did it” defense.
Two enforcement actions in 2025 underscore the operative risk: the Arkansas Supreme Court issued a show-cause order against attorney Dana McClain for hallucinated Microsoft Copilot citations (she resigned), and Judge Timothy L. Brooks (W.D. Ark.) imposed $1,578,172 in fees plus $93,388 in costs on Tony Pirani for ChatGPT-generated fictitious citations.
Bottom line for a 5-50 attorney Arkansas firm: Two binding instruments are in force: Administrative Order No. 25 and the amended Rule 5.3. Active enforcement at both the state Supreme Court and Western District levels makes the risk of sanctions for unverified AI citations concrete and immediate.
Last verified: April 24, 2026