Some courts have, or will, issue standing orders addressing AI. In at least one court in Virginia, lawyers must disclose if they used AI and acknowledge that the lawyer remains responsible for the contact obtained using AI. They do something different: they ask lawyers to certify when they have used AI in preparing a filing and certify that they are responsible for the consequences of filing fictitious cases. A better approach is to require a slight modification of the traditional certificate of service.

When a Virginia lawyer signs a pleading, the lawyer certifies under Code § 8.01-271.1 that it is well-grounded in fact, warranted by existing law, and not filed for an improper purpose. 

So why single out AI? Lawyers do not certify when they use Westlaw or Lexis and both now run on AI. Lawyers do not certify when they search Google, hire a private investigator, or ask a paralegal to pull cases. The decision to use one tool rather than another is classic work product. A rule that forces a lawyer to announce one particular tool, and only that tool, treats an ordinary professional choice as something suspicious.

Picture two briefs filed the same morning. A large firm assigns twenty lawyers, three paralegals, and a contract-attorney service to one brief. No rule requires that firm to disclose any of it. The partner signs, and the court sees a polished product with a single name on it. Across the hall, a solo practitioner uses AI to do the work those twenty people would have done — research, organize, draft, and verify — and the rule requires her to raise her hand and announce it.

All attorneys are responsible for their filings. Yet only the small-firm lawyer has to explain how the sausage was made. AI is an equalizer for small practices; it lets one capable lawyer stand toe-to-toe with an army. A disclosure mandate turns that equalizer into a scarlet letter, and is unnecessary. 

A lawyer who files hallucinated citations should be sanctioned. But the proposed cure is the duty we already carry: read what you file, verify every authority, and own every word above your signature. That duty governs AI exactly as it governs a tired associate at 2 a.m.

Adding another standing disclosure rule feels responsive to a real problem. They are not. They duplicate § 8.01-271.1, they ignore how pervasively AI already runs through everyday legal research, and they lay a unique burden on the lawyers least able to absorb it. Verify your work. Sign your name. That has always been enough and it should be today.