The legal profession is waiting for clarity.
It wants rules. Bright lines. A defined set of dos and don’ts for using artificial intelligence in legal work. Something concrete that says, “This is acceptable,” and “This is not.”
But that’s not how this is going to unfold.
If the last year has taught us anything, it’s that courts are not rushing to create an entirely new regulatory framework for AI. Instead, they are doing something far more consequential. They are applying existing ethical rules to new technology — and in the process, quietly raising the standard.
The question isn’t when courts will “catch up.” It’s whether lawyers understand that, in many ways, they already have.
Consider what’s already in place. The duty of competence. The duty of diligence. The obligation of candor to the tribunal. The requirement to supervise junior attorneys and non-lawyer assistants. None of these rules were written with AI in mind, but all of them apply directly to it.
And courts are making that crystal clear, case by case.
What we are seeing is not a regulatory gap. It’s a regulatory adaptation.
In the wake of high-profile missteps — including the recent filing issues involving Sullivan & Cromwell — judges have not asked for new laws. They have asked a much simpler question: Did the lawyer verify the work before submitting it?
If the answer is no, the analysis is over.
That approach may feel unsatisfying to lawyers who are used to operating within detailed frameworks. But it reflects a deeper reality. The law often evolves through application, not proclamation. And right now, the application is pointing in a very specific direction.
Courts are not interested in regulating the technology itself. They are regulating the conduct of the lawyer using it.
That distinction matters because it shapes what comes next.
There will be more local rules. Some judges have already begun requiring disclosure when AI is used in drafting filings. Others are considering certification requirements, effectively forcing attorneys to affirm that no AI-generated content has been submitted without verification. Bar associations are issuing ethics opinions that emphasize technological competence as part of a lawyer’s baseline obligations.
But none of this is likely to result in a single, universal rulebook.
Instead, we are heading toward a layered system. At the top level, existing ethical duties remain unchanged. Beneath that, courts and jurisdictions will impose procedural guardrails — disclosure requirements, certifications, and, in some cases, sanctions specifically tied to misuse of AI tools.
And beneath all of that, firms themselves will be expected to build internal systems that ensure compliance.
This is where many firms are falling behind.
Policies are being written, but not enforced. Training is being offered, but not internalized. AI tools are being adopted faster than the safeguards designed to manage them. The result is a growing disconnect between what lawyers think is permissible and what courts are actually willing to tolerate.
That gap will close. The only question is how painful the process will be.
For smaller firms, the stakes are particularly high. Large firms can absorb the reputational impact of a single misstep. They have layers of review, internal resources, and the ability to course-correct quickly. Smaller firms do not have that luxury. One error, one sanction, one moment of lost credibility with a court can have lasting consequences.
And yet, the ethical obligations are identical.
There is no sliding scale for responsibility.
What’s coming next is not a dramatic regulatory overhaul. It’s something more incremental, and in many ways, more demanding. Courts will continue to issue orders that clarify expectations. Bar associations will refine guidance. Sanctions will become more targeted and more visible. And over time, a body of precedent will emerge that defines acceptable use of AI in legal practice.
By the time that process is complete, the “rules” will be clear.
They just won’t be written in a single place.
For lawyers looking to get ahead of this curve, the path forward is not particularly complicated. Treat AI as you would any junior contributor. Verify everything. Document your process. Be transparent when necessary. And most importantly, assume that anything you submit will be scrutinized without regard for how it was created.
Because it will be.
The legal profession has always adapted to new tools. From online research databases to e-discovery platforms, each technological shift has required lawyers to recalibrate how they work. AI is no different in that respect.
What is different is the speed.
The rules are not coming someday in the future. They are being built right now, in courtrooms, in sanctions orders, and in the quiet expectations judges bring to every filing they read.
Lawyers who are waiting for permission are going to be disappointed.
Lawyers who understand that the standard has already changed are going to be fine.
Michael J. Epstein, a Harvard Law School graduate, is a trial lawyer and managing partner of The Epstein Law Firm, P.A., a law firm based in New Jersey.
