The Question Nobody Seems to Be Asking
There is a conversation happening across the legal industry about AI adoption. Firms in New York and nationally are racing to integrate AI tools into their practice — contract analysis, legal research, document review, case prediction. The conversation is almost entirely about capability and efficiency. Faster research. Better pattern recognition. Lower costs.
But there is a question that very few attorneys seem to be asking, and it is the one that matters most: if an AI system makes a recommendation, you rely on it, and your client gets harmed, who is liable?
The answer is straightforward. You are.
AI Does Not Carry Malpractice Insurance
The AI system did not sign the engagement letter. It does not have a license to practice law in New York. It does not carry professional liability insurance. If your client sues for malpractice, they are suing you. The fact that you used AI in your process does not change your duty of care, your duty of competence, or your duty of loyalty to the client.
This might seem obvious, but the way many firms are adopting AI suggests they have not fully internalized it. When a firm sets up a workflow where AI does the research and the attorney follows the output — essentially using AI as a decision-making system rather than a decision-support system — they have created a liability problem that their malpractice carrier may not be prepared to cover.
The Critical Distinction: Decision-Making Versus Decision-Support
There is a fundamental difference between using AI as a decision-making system and using it as a decision-support system, and that difference has real consequences for malpractice liability in New York.
A decision-making system is one where the AI produces an output and the attorney acts on it without independent verification. The AI says these are the relevant cases, and the attorney cites them. The AI says this contract clause is acceptable, and the attorney approves it. The AI says the risk profile is low, and the attorney advises the client accordingly. In this model, the attorney has effectively delegated judgment to an algorithm.
A decision-support system works differently. The AI surfaces candidates — potential cases, risk flags, contract issues, research leads. The attorney reviews those candidates independently, applies professional judgment, and makes the call. The AI assisted. The human decided.
If something goes wrong under the first model, the attorney’s defense is weak. You relied on a system you did not fully understand, and you did not exercise independent judgment. That is difficult to defend in a malpractice action in any New York court.
Under the second model, the attorney has a defensible position. You used a research tool. You reviewed its output. You made the judgment call. That is the standard of care — using available tools competently while maintaining professional responsibility for the outcome.
The Ethics Dimension for New York Attorneys
Beyond malpractice liability, there is an ethics question that New York attorneys need to take seriously. The New York Rules of Professional Conduct require competent representation. That includes understanding the tools you use in your practice well enough to evaluate their output.
If you are using an AI system for legal research and you cannot explain how it generated its results — why it selected certain cases, what data it relied on, what its limitations are — you may have a competence problem. Not because using AI is incompetent, but because using any tool you do not understand well enough to verify is a risk to your client.
We have already seen cases nationally where attorneys submitted AI-generated briefs containing fabricated case citations. Those situations were not caused by AI being unreliable. They were caused by attorneys treating AI output as verified without doing their own review. The tool did what it does. The attorney failed to do what attorneys must do.
How We Handle AI Liability at Travis & DeBlase PLLC
At our firm, every high-stakes decision that goes to a client gets human verification and independent judgment. AI surfaces options. Humans choose. That is not a tagline — it is an operational rule built into our workflow.
When our AI system flags a contract risk, the attorney independently reviews the clause, checks the underlying authority, and makes the recommendation. When the system generates a research memo, the attorney reads the cited cases before relying on them. When the system predicts a litigation outcome, we treat that prediction as one data point among many, not as the answer.
Is this slower than pure AI automation? Yes. But it means we can stand behind every recommendation we make to our clients. And if something goes wrong, our position is clear: we used the best available tools, and we exercised professional judgment at every decision point.
What Clients Should Ask Their New York Attorney About AI
If you are a business hiring outside counsel in New York, the AI liability question should be part of your due diligence. Ask your attorney directly: when you use AI in your practice, how do you verify its output? What is your process for ensuring that AI-generated work product meets the standard of care? Who is responsible when the AI is wrong?
The right answer is: the attorney is always responsible, and the firm has a verification process to ensure that AI output gets independent human review before it reaches the client. If your counsel cannot articulate that clearly, you may want to understand more about how they are using these tools on your matters.
AI Done Right Protects Your Interests
Travis & DeBlase PLLC uses AI to deliver better, faster, and more cost-effective legal services to businesses in New York City. But we never lose sight of the principle that matters most: the attorney’s judgment is the product, and AI is the tool that supports it. If you are looking for a Manhattan law firm that takes both innovation and professional responsibility seriously, contact us to discuss how we can help.