On one side, it is said that AI will transform legal practice beyond recognition and lawyers should feel a sense of existential dread. The other side says the effects of AI are completely overhyped and traditional methods will persist indefinitely. Both positions are presented with far more confidence than the evidence warrants.
What is actually happening is more nuanced, and more useful to think about carefully.
Certain parts of legal work will always and irreplaceably require professional judgment. Advising a client on how to navigate a complex commercial dispute. Assessing litigation risk in circumstances where the law is unsettled. Making a call about strategy that depends on understanding the client's business, their appetite for risk, and the practical realities of the matter. That judgment depends on experience, relationships, contextual understanding, and accountability to a client. It cannot foreseeably be automated, and there are good structural reasons why the legal system would not want it to be.
Other parts of legal work are logistical. Finding relevant authorities across multiple databases. Organising research material into a usable structure. Checking whether a legal proposition is actually supported by the source cited for it. Extracting a specific passage or principle from a lengthy judgment. Identifying whether recent decisions have shifted the landscape on a particular point.
These tasks require care and accuracy. They are important. But they do not require the kind of judgment that takes years of practice to develop. They require patience, diligence, and time — and they consume a great deal of all three.
A junior solicitor who spends four hours locating and organising authorities for a research memorandum, when there are better methodologies toinform research and analysis, is not spending those four hours developing the analytical skills that will make them a better lawyer. A senior associate who spends time verifying that the cases cited in a memo actually say what the memo claims they say is not spending that time on the higher-order work their experience equips them to do. A partner who could be refining strategy or deepening a client relationship is instead reviewing work product for completeness at a level that sits well below their capability.
This pattern is familiar to every practitioner in Australia. It is accepted as a normal cost of legal work. But the fact that something is normal does not mean it is efficient, and it does not mean it should be permanent.
The current allocation of lawyer time to mundane logistical tasks does not do clients or practitioners any favours. Clients pay for hours that include a substantial component of mechanical work. Practitioners spend portions of their careers on tasks that do not reflect or develop their professional skill. Firms absorb inefficiencies that compound across every matter and every lawyer.
The productive question is not whether AI will replace lawyers. It will not, for the same reason that legal work has always depended on human judgment exercised in circumstances of genuine complexity and consequence. Courts require accountability. Clients require trust. The law requires interpretation that is sensitive to context in ways that no current or foreseeable technology can replicate.
The productive question is which parts of legal work are genuinely amenable to automation, and what happens when those parts are handled more efficiently.
If a tool can surface relevant authorities with greater reliability and speed than manual keyword searching, that is worth adopting — not because it replaces the lawyer's role, but because it frees the lawyer to spend more time on the work that actually requires legal skill. If a tool can verify citations or flag inconsistencies in a research memo before it reaches a reviewer, that saves senior time for the supervision that matters: assessing the quality of the analysis, the strength of the argument, and the soundness of the advice.
The firms and practitioners who will benefit most from AI in legal practice are those who can distinguish clearly between the work that requires their judgment and the work that merely requires their time. Automating the latter protects the former. It gives lawyers more room to do the work they trained for, and it gives clients a service that is more focused on the things that actually affect outcomes.
That distinction — between judgment and logistics — is where the real conversation should be. The binary of existential threat versus irrelevant hype is a distraction from it.
