After a Client Asks "Did AI Write This?", How NJ Attorneys Should Handle Disclosure Under RPC 1.4

A client emails back thirty minutes after receiving a draft settlement agreement. The message is short: "Did a computer write this?"

That question is landing in more NJ attorneys' inboxes than ever, and most solo and small-firm lawyers don't have a prepared answer. Some deflect. Some give a vague "I reviewed everything carefully." Some panic and overcorrect by disclaiming the tool entirely, then wonder whether they've just implied incompetence.

None of those responses is great. And under New Jersey RPC 1.4, the ethics stakes here are more concrete than most practitioners realize.

What RPC 1.4 Actually Obligates You to Do

RPC 1.4 requires attorneys to keep clients reasonably informed about the status of a matter and to promptly respond to reasonable requests for information. The rule also requires explaining matters to the extent reasonably necessary for the client to make informed decisions about the representation.

That last clause is where AI disclosure gets interesting. The question isn't whether you must announce every tool you use (you don't disclose which word processor you type in). The question is whether the client's ability to make an informed decision about the work product is affected by how it was generated.

For a routine form letter, probably not. For a contract or motion that the client will sign or file, the analysis shifts. If a client asks directly, RPC 1.4(b) almost certainly requires a truthful, substantive answer. Saying "I reviewed it" when the client is asking about origin, not review, skirts the line.

The New Jersey Supreme Court's Committee on the Unauthorized Practice of Law and the ACPE haven't issued a formal opinion specifically on AI disclosure as of mid-2025, but the existing RPC 1.4 framework maps cleanly onto the situation. Other state bars, including California and Florida, have published guidance treating AI use as a material fact clients may reasonably want to know. NJ attorneys shouldn't wait for a formal opinion to treat this as a live obligation.

The Disclosure Gap Most Firms Have Right Now

Here's the practical problem: most NJ solo attorneys using AI tools (Clio Draft, CoCounsel, ChatGPT, Copilot, or otherwise) have no written protocol for what to tell clients about AI involvement. The retainer agreement says nothing about it. There's no standard language in transmittal emails. The attorney decides in real time, call by call, how much to share.

That inconsistency is itself a risk. If you tell one client "yes, I used AI and here's how I reviewed it" and another client gets a vague non-answer, you've created an uneven standard that a disciplinary panel could use to question your judgment on the second matter.

Building a Disclosure Protocol That Actually Works

The goal isn't to confess every tool use unprompted, that would be impractical and would actually confuse most clients. The goal is a consistent, thought-through approach that satisfies RPC 1.4 and builds trust rather than eroding it.

In your engagement letter or retainer agreement, add a short paragraph. Something like: "This firm may use AI-assisted drafting tools in preparing documents for your matter. All AI-generated content is reviewed and edited by the supervising attorney before delivery. You may ask at any time about the tools used in your representation."

One paragraph. It normalizes the conversation before the client has to initiate it.

In transmittal emails for AI-assisted documents, a brief note works well: "Attached is the draft agreement. This was prepared with the assistance of AI drafting tools and reviewed by me before sending. Please read it carefully and flag any sections you'd like to discuss." That's not an apology, it's professional transparency.

When a client asks directly, answer directly. "Yes, I used [tool] to generate an initial draft, then reviewed and revised it for your matter. Here's what I changed and why." That response demonstrates competence, not weakness. Clients who ask the question are often sophisticated enough to appreciate the honest answer far more than a hedge.

The Informed Consent Question

A few state bars have gone further, suggesting that using AI on client matters may require explicit informed consent under certain circumstances, particularly when the tool involves third-party processing of confidential data. NJ's RPC 1.6 confidentiality angle has been covered elsewhere, but it's worth flagging here: if your AI platform sends client content to a third-party server under terms that don't include a data processing agreement, the disclosure obligation under RPC 1.4 may intersect with a confidentiality problem under RPC 1.6. The two rules aren't isolated from each other.

If you're using a consumer-tier AI product without a BAA or DPA, disclosing AI use may actually surface a deeper issue you need to address first.

Start With the Retainer

If you do one thing this week, open the retainer agreement template you send to every new NJ client and add the AI disclosure paragraph described above. That single change converts an ad hoc, inconsistent practice into a documented, firm-wide standard, which is exactly what a disciplinary body would want to see if a client complaint ever arose about AI-assisted work.

The client who emails asking "did a computer write this?" deserves a real answer. Having that answer ready before the question arrives is what RPC 1.4 is pointing you toward.