Hawaii

Does Hawaii require attorneys to disclose AI use in court filings? No — and the state made a deliberate, documented decision to keep it that way. Hawaii's Committee on AI in the Courts issued its Final Report in December 2025, concluding that existing rules are enough. No new mandatory disclosure requirements were recommended.

What makes Hawaii's situation interesting is why the review happened at all. Multiple complaints about attorneys citing fake AI-generated case law prompted the committee's formation. Hawaii looked at the problem, studied it thoroughly, and concluded the answer isn't new rules — it's better enforcement of existing ones.

The Committee's Final Report and Its Conclusions

The Hawaii Supreme Court's Committee on AI in the Courts released its Final Report in December 2025 after months of study. The committee's core finding: existing Rules of Professional Conduct and Rules of Court adequately address the risks posed by AI in legal practice. The committee explicitly considered and rejected mandatory AI disclosure requirements for attorneys. Instead, it recommended enhanced education, better awareness of AI risks among practitioners, and more vigorous enforcement of existing candor and competence obligations.

The Fake Cases That Triggered the Review

Hawaii's AI review wasn't academic — it was reactive. Multiple instances of attorneys citing AI-fabricated case law in Hawaii courts prompted the Supreme Court to form the committee. These incidents involved attorneys who used AI tools to generate legal research without verifying citations, resulting in filings that referenced cases that didn't exist. The committee studied these incidents and concluded that the problem wasn't a gap in the rules — it was a failure to follow rules that already existed. Attorneys who cite fake cases are violating their duty of candor regardless of whether the fake citations came from AI or their own imagination.

What 'Existing Rules Are Enough' Means in Practice

Hawaii's position means the same Rules of Professional Conduct that governed practice before AI continue to govern it now. Rule 1.1 (competence) means attorneys must understand AI tools well enough to identify their limitations. Rule 3.3 (candor) means every citation must be verified before submission. Rule 1.6 (confidentiality) means client data can't be fed into AI tools without appropriate safeguards. Rule 5.3 (supervision) means attorneys are responsible for AI outputs the same way they're responsible for paralegal work. No new rule is needed because these obligations are technology-neutral.

How Hawaii Differs from Federal Courts in the State

The District of Hawaii — the state's sole federal court — may have different requirements than state courts. Federal judges in Hawaii can and do issue individual standing orders on AI use in filings. These orders may require disclosure, certification, or specific representations about AI involvement. Hawaii state courts, following the committee's recommendations, impose no such requirements. For attorneys practicing in both forums on the islands, this means checking federal standing orders while operating freely under state court rules.

Lessons from Hawaii's Approach

Hawaii's deliberate decision offers a useful framework for firms anywhere. First, the absence of a disclosure rule doesn't mean the absence of obligations — existing professional conduct rules are fully enforceable against AI-related errors. Second, education matters more than rulemaking. The committee emphasized that attorneys need to understand AI risks, not just check compliance boxes. Third, the incidents that triggered Hawaii's review are a warning: real attorneys in real courts submitted real filings with fake cases. Your firm's verification protocol is your first line of defense, regardless of what state you practice in.

The Bottom Line: Hawaii's December 2025 committee report deliberately concluded that existing rules are sufficient — no mandatory AI disclosure was recommended, but the fake citation incidents that triggered the review show why rigorous verification matters regardless.