Competence Has a New Baseline — What NJ RPC 1.1 Actually Demands From Solo Attorneys Using AI in 2025

When NJ RPC 1.1 was amended to include the duty to keep abreast of changes in the law and its practice, most attorneys read the word "practice" and thought: CLE credits, updated court rules, maybe a new filing portal. Few read it as a standing obligation to understand the AI tools increasingly embedded in how legal work gets done.

That reading is no longer defensible.

In 2025, the competence floor for New Jersey practitioners has quietly shifted. Not because the Bar issued a sweeping new opinion. Not because a judge sanctioned someone by name. But because the underlying factual reality — what a reasonably competent attorney knows and uses — has changed. And courts, clients, and disciplinary panels tend to calibrate expectations against that reality, not the one from five years ago.

What Comment 8 to RPC 1.1 Actually Says

The New Jersey Rules of Professional Conduct, like the ABA Model Rules, include Comment 8 to Rule 1.1, which specifies that competence includes "keeping abreast of changes in the law and its practice, including the benefits and risks associated with relevant technology."

That phrase — benefits and risks — is doing more work than most practitioners appreciate. It doesn't say you must use AI. It says you must understand it well enough to make an informed decision about whether and how to use it. Ignorance is not a safe harbor; it's the problem Comment 8 is designed to address.

For a NJ solo attorney in 2025, this has concrete implications:

• Document drafting: If AI-assisted drafting is now standard practice for a given document type in your practice area, producing work product without understanding those tools — and without checking their output — is a competence question, not just an efficiency one.
• Legal research: If you're using a general-purpose LLM for case law research without understanding its hallucination risk, and you serve a client based on a fabricated citation, RPC 1.1 is the first rule cited in any disciplinary complaint — not 3.3.
• Case assessment: AI tools are increasingly used for early case evaluation in litigation, family law, and PI practices. Failing to understand what those tools surface — or suppress — affects the quality of the advice you give.

The Competence Gap Nobody Wants to Measure

Here's the uncomfortable truth: most solo practitioners in New Jersey are using at least one AI-adjacent tool right now — whether it's a smart form engine in their case management software, a generative AI assistant in Microsoft Word, or a standalone tool like Clio Draft or Harvey. And most of them have received no formal training on how those tools work, what they get wrong, or when to distrust their output.

That's not a criticism. The legal technology market moved faster than any reasonable professional development calendar could track. But the gap is real, and it's widening.

RPC 1.1 doesn't require you to become a machine learning engineer. It requires you to understand, at a functional level:

1. What the tool is doing — Is it retrieving from a live database, or generating text probabilistically from training data? The distinction matters enormously for research and drafting tasks.
2. What it gets wrong, and why — Hallucinations in legal AI aren't random glitches. They cluster around obscure statutes, older case law, jurisdiction-specific procedural rules, and prompts that lack sufficient context. NJ practitioners should be particularly aware that tools trained primarily on federal or large-jurisdiction data often underperform on New Jersey-specific matters.
3. What your verification workflow is — Competence in AI-assisted work is not about the tool. It's about what you do after the tool runs. Do you cite-check every case independently? Do you review the full document against the client's facts? If the answer is "usually," that's a liability problem.

A Practical Competence Baseline for NJ Solo Attorneys

Adopt these as non-negotiable minimums before relying on any AI tool in client matters:

• Run a hallucination stress test before trusting any legal AI tool for research. Ask it to cite a case you know doesn't exist. Ask it for a statute section in an obscure area. See how it fails, so you know what to watch for.
• Document your verification steps in your file notes. If a complaint is ever filed, your ability to show a deliberate review process is the difference between a reprimand and a dismissal.
• Know the training data cutoff of every tool you use. For anything time-sensitive — recent NJ appellate decisions, new court rules, updated agency regulations — assume the tool doesn't know, and verify independently.
• Build a short AI tool log for your practice. For each tool you use, note what it's used for, what it's not trusted to do, and who reviews its output. This is the foundation of any AI policy, but more importantly, it forces you to think clearly about where you're relying on automation and where you're maintaining independent judgment.

The Disciplinary Horizon

The New Jersey Office of Attorney Ethics has not yet issued AI-specific guidance that goes beyond the general principles embedded in the RPCs. But ABA Formal Opinion 512 — which NJ attorneys are expected to consider as persuasive authority — made clear that competence, supervision, and confidentiality are the three pillars of responsible AI use, and competence is foundational to the other two.

It is not a question of whether the disciplinary apparatus will catch up to AI-related misconduct. It is a question of when, and which practitioners get to serve as the case studies.

Solo attorneys are disproportionately exposed. No firm IT department. No general counsel. No senior partner reviewing the work. The competence obligation doesn't scale down because the practice is small — and in some respects, the accountability is sharper precisely because the supervision chain is shorter.

Understanding what RPC 1.1 demands in 2025 isn't an academic exercise. It's the due diligence of staying in practice.