Disciplinary Exposure Under NJ's Preliminary AI Guidelines: Is Any of It Actually Enforceable Right Now?

The NJ Supreme Court issued its "Preliminary Guidelines on New Jersey Lawyers' Use of Artificial Intelligence" on January 24, 2024. Most coverage has focused on what the guidelines recommend. Far less attention has gone to a harder question: what happens to a NJ attorney who doesn't follow them?

That question matters more than it might seem, especially for solos and small firms that are making real decisions right now about which AI tools to use, how much to rely on them, and what to tell clients.

What the guidelines actually are

The word "preliminary" in the title isn't decorative. The NJ Supreme Court used it deliberately to signal that this document is a starting point, not a finished regulatory framework. The guidelines themselves acknowledge they are issued in anticipation of further development as AI technology and legal practice evolve.

That framing has a practical consequence. The guidelines do not amend the New Jersey Rules of Professional Conduct. They do not create new, free-standing duties with their own enforcement mechanisms. What they do is map AI-related conduct onto existing RPC obligations, primarily around competence, confidentiality, supervision, and candor. If a lawyer faces discipline for AI-related conduct in New Jersey today, the charge will be brought under those existing RPCs, not under the guidelines themselves.

This is not a loophole. It's how NJ court guidance typically works. The Supreme Court issues guidance; the RPCs provide the teeth.

Where the real exposure sits

Because discipline runs through the RPCs, the practical question for a solo attorney is which rules are most likely to be implicated by AI missteps. The guidelines point clearly at a few.

Supervision is one. NJ RPC 5.3(b) requires a lawyer who has direct supervisory authority over a non-lawyer to make reasonable efforts to ensure that person's conduct is compatible with the lawyer's professional obligations. The guidelines extend this logic to AI tools: a lawyer using AI is expected to review AI-generated output before it goes to a client or a court. A solo who submits a brief containing hallucinated citations without reviewing the AI output isn't just careless. Under the guidelines' framing, they've arguably failed a supervision-adjacent duty already embedded in the RPCs.

Candor toward the tribunal is another. NJ RPC 3.3(a)(1) prohibits a lawyer from making a false statement of fact or law to a tribunal. An AI-generated citation to a case that doesn't exist is a false statement. The guidelines don't create that rule; the rule already exists. The guidelines just make plain that AI doesn't excuse you from it.

The enforcement gap, and why it's not a safe harbor

Here's the uncomfortable part: because the guidelines are preliminary and non-binding on their face, some attorneys treat them as optional. That's a mistake, and the logic is worth unpacking.

Even if the guidelines carry no independent disciplinary weight, they do something important: they put the bar on notice. If the NJ Office of Attorney Ethics investigates an AI-related complaint after January 2024, those guidelines will almost certainly be part of the factual backdrop. A disciplinary body looking at whether a lawyer exercised reasonable competence or adequate supervision will have a publicly available document, issued by the same Supreme Court that oversees discipline, describing what reasonable AI use looks like.

That's not the same as binding authority, but it's not nothing. Courts and disciplinary tribunals regularly look to official guidance when evaluating whether conduct met a professional standard, even when the guidance itself isn't a rule. The guidelines create a reference point that didn't exist before January 2024.

What the ACPE has and hasn't said

The Advisory Committee on Professional Ethics has not issued an AI-specific ethics opinion as of this writing. The most relevant ACPE opinion touching on technology is ACPE Opinion 701 (2006), which addresses the duties attorneys owe when storing client information electronically, including cloud storage. That opinion roots its analysis in NJ RPC 1.6 confidentiality obligations. It does not address generative AI specifically.

Attorneys looking for ACPE-level guidance on AI tools will not find a dedicated opinion yet. The Supreme Court's preliminary guidelines are currently the closest thing New Jersey has to official, court-level direction on the subject.

A practical frame for today

Think of the current moment as a transition period with real stakes. The guidelines are preliminary, but the RPCs they reference are not. Discipline is possible now, under existing rules, for AI-related conduct that violates confidentiality, candor, or supervision obligations.

The most defensible position for a NJ solo attorney is to treat the guidelines as the Court's stated view of what the existing RPCs require in an AI context, not as a separate, optional document. That means reviewing AI output before it leaves your desk, not feeding client confidences into AI tools without understanding where that data goes, and documenting your review process in matters where AI played a material role.

A formal AI policy helps with that documentation, but even informal habits like keeping a short note in your file about which tools you used and what you reviewed will matter if your conduct is ever questioned.

The guidelines will eventually be replaced by something more permanent. When that happens, the transition from "preliminary" to "final" may narrow the interpretive gap considerably. Attorneys who've built their practice around the guidelines' substance now will have less adjusting to do then.

What the rule requires

NJ RPC 3.3(a)(1) prohibits making a false statement of fact or law to a tribunal. NJ RPC 5.3(b) requires lawyers with direct supervisory authority over non-lawyers to make reasonable efforts to ensure conduct compatible with professional obligations. The NJ Supreme Court's Preliminary Guidelines on New Jersey Lawyers' Use of Artificial Intelligence (January 24, 2024) map these and other existing RPC duties onto AI use but do not themselves create new, independently enforceable obligations separate from the RPCs. ACPE Opinion 701 (2006) requires attorneys to take reasonable steps to protect client confidentiality when storing client information electronically; it does not address generative AI.

What I recommend

Treat the January 2024 preliminary guidelines as the NJ Supreme Court's public statement of what the existing RPCs require in an AI context. Even though they aren't independently enforceable, they will almost certainly inform how any AI-related disciplinary complaint is evaluated. Keep a short contemporaneous note in your file for matters where AI was used in a meaningful way: what tool, what task, and what review you performed. That note costs you nothing to create and gives you something concrete to point to if your judgment is ever questioned.