When the NJ Supreme Court Calls AI Use "Ongoing": What That Single Word Actually Means for Solo Attorneys

Most NJ attorneys who read the Supreme Court's "Preliminary Guidelines on New Jersey Lawyers' Use of Artificial Intelligence" (January 24, 2024) did so once, maybe twice, and moved on. That's understandable. The document is brief, and at first read it sounds like a restatement of things already covered by the RPCs: be competent, protect client data, don't mislead the court.

But there's a phrase buried in the guidelines that hasn't gotten nearly enough attention, and it has real operational consequences for a solo practitioner. The guidelines describe the duty to maintain current knowledge of AI tools as an ongoing obligation, not a one-time assessment. That word does a lot of work.

What "ongoing" actually signals

The guidelines tie lawyer AI obligations to NJ RPC 1.1, which requires competent representation, including the legal knowledge, skill, thoroughness, and preparation reasonably necessary for the representation. The guidelines extend this to the tools lawyers use, and they do so in forward-looking terms.

That framing matters because most solos approach AI adoption the way they approach buying a new printer: evaluate it once, set it up, and forget it. An AI tool is not a printer. The underlying models that power legal AI products are updated, sometimes substantially, without much fanfare. A product you evaluated in February 2024 may behave differently in October 2025. Its outputs on the same prompt type may have changed. Its accuracy on legal citation retrieval may have improved or degraded. Its terms of service may have been quietly amended.

Under the guidelines' "ongoing" framing, a solo attorney who ran a competent evaluation of a tool at the point of adoption and then stopped monitoring it has, at minimum, a gap in their professional workflow, even if they haven't yet made an error.

The malpractice angle no one is talking about

Ethics guidelines aren't malpractice rules, and the NJ Supreme Court's AI guidelines specifically don't carry the direct enforcement weight of the RPCs. But they inform what "reasonable" looks like when a disciplinary body or a malpractice plaintiff reconstructs your decision-making.

Consider a hypothetical: a solo family law attorney started using an AI drafting tool in early 2024 after reviewing its outputs against manual work and feeling satisfied with accuracy. Eighteen months later, the attorney submits a motion that contains a plausible-sounding but incorrect statutory reference, generated by the same tool, which the attorney didn't independently verify because the tool had always been reliable before. The guideline's "ongoing" competence standard is the lens through which that attorney's conduct gets evaluated. Did they keep up with the tool? Did they know the model had been updated? Did they have a verification routine in place?

None of that requires the attorney to be a software engineer. It requires them to treat AI outputs with the same professional skepticism they'd apply to research from a first-year law clerk.

What this looks like practically for a one-person shop

The guidelines don't specify a review cadence, and frankly neither does any other NJ authority at this point. The ACPE has not issued an AI-specific opinion. The January 2024 guidelines remain the primary NJ-specific touchpoint. So the "ongoing" standard has to be self-defined.

A solo practitioner can build a defensible routine without much overhead. A quarterly spot-check, where you run the same three or four test prompts you used during initial evaluation and compare outputs, takes under an hour. Subscribing to a vendor's changelog or update notifications costs nothing. Keeping a short internal note (a one-paragraph entry in a practice management file) documenting when you last reviewed a tool and what you found creates a record that matters if your judgment is ever questioned.

The guidelines also emphasize that lawyers remain responsible for all AI-assisted work product. That's not a throwaway line. It means the "ongoing" obligation isn't just about the tool, it's about maintaining your own independent judgment about when to trust it, when to double-check it, and when to set it aside entirely.

What the rule requires

The NJ Supreme Court's Preliminary Guidelines on New Jersey Lawyers' Use of Artificial Intelligence (January 24, 2024) state that attorneys must maintain competence in the AI tools they use, consistent with the obligations under NJ RPC 1.1. The guidelines characterize this as a continuing duty, not a one-time assessment. Attorneys remain responsible for all work product, including work generated with AI assistance. The guidelines do not set a specific review schedule or prescribe particular monitoring methods.

What I recommend

Set a calendar reminder every 90 days to run a short output-quality check on any AI tool you use for client-facing work. Keep a one-paragraph log entry each time you do it, noting the date, what you tested, and whether anything changed. Sign up for product update notifications from your AI vendor. If a tool's underlying model gets updated and you haven't re-evaluated it, treat that update the same way you'd treat hiring a new law clerk: verify the work more closely until you've rebuilt your baseline confidence. That's not paranoia, it's what the guidelines' "ongoing" language points toward, and it's the kind of documented diligence that holds up under scrutiny.