There’s a question I would ask my own lawyer in 2026: how do you use AI in your work?
Not “have you tried ChatGPT.” But has AI become part of how you research, review, analyse and structure.
A couple of years ago the question would have been premature. Today, the answer tells me a great deal about a lawyer.
Let me be clear: I don’t think AI will replace lawyers. I think the opposite. But I expect a colleague to know how to use the technology available to them — and to know exactly where it stops being reliable.
A lawyer’s value is no longer measured in hours
For decades, the legal industry was built around time. Research takes time. Contract review takes time. Due diligence takes time. Reconciling hundreds of pages takes time.
AI breaks that equation.
I see it in my own work. When we ran a licence-compliance audit on a product — 137 components, each with its own licence and its own restrictions — part of the mechanical reconciliation took a fraction of the time it would have by hand. But the machine didn’t decide for me which blockers were critical or how to close them. And it wasn’t the one signing the report to the investor.
Which raises an uncomfortable question for the profession: if technology makes part of the work faster, what exactly is the client paying the lawyer for?
My answer is judgment.
For the ability to see the risk, recognise what actually matters here, challenge someone else’s assumption, design the transaction, and decide where there is no obvious answer.
Not for five hours of manual work that could honestly have been done in one.
AI should make a lawyer think more, not less
There’s a misconception that using AI means handing your thinking to a machine. Done properly, it’s precisely the reverse.
AI removes the mechanics and gives the lawyer back time for the thing they were hired for.
Take my own field — intellectual property. AI will help organise a portfolio, compare wording, speed up preliminary research. But it won’t answer the central question: does this company actually own the technology it believes it owns?
To answer that, you have to reconstruct years of relationships between founders, developers, contractors and legal entities across several countries. You have to build the ownership chain — as I did for InterMIND between the Estonian and Emirati companies. That isn’t document processing. That’s architecture, and a human is accountable for it.
The danger isn’t AI. It’s using AI uncritically
There’s another side to this.
A lawyer who blindly trusts AI is more dangerous than one who doesn’t use it at all.
Generative models get things wrong. They can invent a provision, confuse jurisdictions, and deliver an incorrect conclusion with total confidence. Then there’s a whole separate layer: client confidentiality and data protection.
So I have a rule. Whatever AI produces is a first layer of analysis, never the final authority. Legislation I verify against the primary source. A regulator’s position I confirm. Case law I validate. Confidential data doesn’t go into open models.
The principle is simple: AI can assist with the work. It cannot assume the lawyer’s responsibility.
Clients have become more sophisticated — and that’s a good thing
My clients use AI themselves. Founders arrive at meetings having already worked through the basic requirements. Executives run a contract through a model before sending it to me.
That changes the relationship. The lawyer is no longer the person in the room with a monopoly on information.
Information is now abundant. Judgment is what’s scarce.
Increasingly, clients don’t want the rule recited — they want to know what it means for this particular deal. They expect speed. And they ask questions about inefficient processes. For the profession, that’s healthy.
And there’s a new layer of legal risk
AI is changing not only how lawyers work, but what lawyers are obliged to understand.
Companies are embedding AI into products, into support, into hiring, into decision-making. Which raises the questions: who owns a model’s output, what licences was the training data under, who is liable for the algorithm’s decision, and how does all of this sit with GDPR and local requirements.
I’m General Counsel at an AI platform, and I see these questions from the inside every day.
So for a lawyer advising technology businesses, “I don’t really use AI” is getting harder to distinguish from “I don’t fully understand how my client’s business works.”
You don’t need to be an engineer. But you do need to understand the technology well enough to see where the risk actually sits.
So should you refuse a lawyer who doesn’t use AI?
Not automatically.
Experience, integrity and professional judgment still matter far more than familiarity with any given technology. A strong lawyer doesn’t become obsolete for preferring traditional methods.
But ask the questions. How do you use technology? How do you verify what the AI produced? How do you protect my data? What’s automated, and what stays under human review?
And above all: is the technology making your advice better — or have you simply not noticed that the profession has changed?
The best lawyers of the next decade won’t be the ones who outsourced their thinking to a machine. Nor the ones pretending it doesn’t exist.
They’ll be the ones who know exactly what the machine can do, exactly what it can’t, and where human judgment still decides everything.
That, to me, is where the future value of our profession lies.
Anastasia Khurtina is a UAE-based corporate lawyer specialising in intellectual property, technology law, corporate structuring and cross-border matters.
