The Truth About Client Communication Disclosures When NJ Attorneys Use AI to Draft Letters and Emails

Most NJ attorneys who use AI tools day-to-day have quietly settled on a comfortable assumption: if the work product is accurate and you reviewed it, the client doesn't need to know how the sausage got made. That's a defensible position in some contexts. In the context of client communications specifically, it's worth pressure-testing.

RPC 1.4 doesn't get much airtime in the AI ethics conversation. Confidentiality, competence, and supervision dominate. But RPC 1.4 sits underneath all of it. It requires a New Jersey attorney to keep a client "reasonably informed about the status of a matter" and to "promptly comply with reasonable requests for information." More quietly, it also requires attorneys to "explain a matter to the extent reasonably necessary to permit the client to make informed decisions regarding the representation."

That last clause is where AI-assisted client communications start to get complicated.

What RPC 1.4 Is Actually Getting At

The rule isn't about transparency for its own sake. It's functional: clients need enough information to exercise their own judgment about their case and their attorney. When AI drafts the letter informing a client of a settlement offer, or the email walking a client through their litigation options, and that draft goes out after only a cursory read, the client is relying on something they don't know exists.

That may not be a violation in most cases. But it creates a category of risk that few NJ solo practitioners have formally thought through.

Consider a scenario: a client receives an AI-drafted email from their attorney explaining three options for responding to a demand. The explanation is technically accurate but omits a nuance the attorney would have caught in a slower, manual drafting process. The client picks the wrong option partly because the explanation wasn't quite right. Whose error was that? The attorney reviewed and sent it, so the attorney owns it. But the fact that AI generated the first draft, and that the review was cursory, is now relevant context in any malpractice or grievance proceeding.

The question of disclosure isn't purely philosophical. It's about audit trails and professional accountability.

Where the NJ Bar Has and Hasn't Spoken

The New Jersey Supreme Court and the ACPE (Advisory Committee on Professional Ethics) have not issued a formal opinion specifically addressing AI-generated client communications as of mid-2025. The ABA's Formal Opinion 512 touched on confidentiality and competence in AI use, but it didn't squarely address whether attorneys must disclose AI assistance in the substance of client-facing communications.

That silence cuts two ways. It means there's no bright-line rule requiring disclosure, which gives NJ practitioners flexibility. It also means there's no safe harbor to stand behind if a grievance examiner concludes that an attorney's AI-generated communications fell short of what RPC 1.4 demands.

The practical reality is that ethics guidance on AI almost always lags actual practice by two to three years. NJ attorneys adopting AI-assisted communication workflows now are operating in that gap.

A More Practical Frame: What Should You Disclose, and When?

Blanket disclosure on every email is probably not necessary and may actually confuse or alarm clients unnecessarily. What's worth considering instead is a tiered approach.

For routine status updates and scheduling confirmations drafted with AI assistance, disclosure is arguably overkill. The stakes are low and the content is verifiable at a glance.

For substantive communications explaining legal options, risks, or recommendations, the standard is higher. If an AI-drafted explanation shapes a client's decision in a material way, the attorney's review should be rigorous enough that the attorney can genuinely own the analysis. If you can't say with confidence that you'd have written essentially the same thing yourself, you've got a supervision problem more than a disclosure problem.

For initial engagement communications, fee explanations, or letters that document advice given, disclosure of AI assistance in your retainer or engagement letter is a clean solution. A single sentence in your standard engagement paperwork, something like "Our firm uses AI-assisted tools to improve efficiency, and all communications remain subject to attorney review and supervision," covers the field without making every client email a disclosure event.

The Retainer Clause Most NJ Solos Are Missing

The simplest fix available to any NJ solo attorney right now is to update the engagement letter. Add one sentence. Keep it plain. This does several things at once: it satisfies any future interpretation of RPC 1.4's informed-consent dimension, it preempts client complaints about "impersonal" AI-generated letters, and it gives you a documented basis for saying you ran a transparent practice.

Draft that clause this week, before the ACPE issues formal guidance that forces you to retrofit it into every active matter.