Last summer, I was interning at a corporate law firm. One afternoon, a senior associate needed to review a stack of contracts for red flags before a closing. I remember thinking this was going to eat up someone's entire afternoon. Instead, she ran the documents through an AI tool, skimmed the flagged clauses in about fifteen minutes, and moved straight on to her next task.
I just stood there doing the math in my head. That was exactly the kind of work junior lawyers are usually handed to learn the ropes. If a machine could handle that first pass before lunch, what was I spending years of law school training to do?
I did not say anything at the time. But the question stayed with me for weeks.
As a fifth-year law student who has since started using AI tools for research, drafting, moot memorials, and legal notices during internships, I have landed on an answer that is more nuanced, and honestly more reassuring, than the panic I felt that afternoon.
No, AI will not make lawyers obsolete. But it will make some old ways of working obsolete. More importantly, it will raise the standard for what good lawyering looks like.
That distinction matters. The future of law is not a contest between lawyers and machines. It is a shift from doing every repetitive task manually to using better tools while keeping human judgment, accountability, and strategy at the centre of the work.
My Own Experience Using AI as a Law Student
Since that internship, I have used tools like Gemini and Lexi for legal research and drafting, mostly for moot court competitions, memorials, and legal notices during my internships.
My first attempts were rough. I did not know how to prompt well, and the drafts I got back were basic, generic, and sometimes inaccurate. Often, the output needed so many changes that it felt faster to write from scratch. I remember thinking the tools were overhyped.
But once I learned how to give better instructions, everything changed. When I started adding context, jurisdiction, facts, procedural posture, and the exact issue I wanted the tool to address, the same tools became genuinely useful. They helped me build first drafts faster, organize research more clearly, and find angles I may not have considered immediately.
This taught me something I did not expect: the gap between 'AI is useless' and 'AI is a game-changer' is often not the technology. It is the skill of the person using it.
That is where a legal AI tool like Lexi fits naturally. Its value is not that it removes the lawyer from the process. Its value is that it can support the first pass: research to verify, clauses to review, drafts to refine, and citations to check. The lawyer still has to direct the work, test the output, and take responsibility for the final answer.
Why Obsolescence Is the Wrong Fear
The word 'obsolete' makes it sound as if the entire legal profession is about to disappear. That is not what is happening. What is becoming less defensible is the idea that lawyers should spend hours on repetitive work simply because that is how the profession has always done it.
There is a difference between legal tasks and the legal profession. AI can assist with many tasks that sit at the surface layer of legal work:
· scanning contracts for red flags and inconsistencies;
· summarizing long judgments, pleadings, or case files;
· creating first drafts of notices, clauses, memos, and research notes;
· organizing large volumes of information into a clearer structure.
But the legal profession is not just the production of text. It is the exercise of judgment under uncertainty. It is knowing which argument matters, which point to leave out, how a judge may read a fact pattern, how a client understands risk, and when a technically correct answer is not the most practical one.
AI can reduce the time spent on repetitive research and first-draft work. It cannot become the professional who signs the filing, faces the client, appears before the court, or carries ethical accountability for the advice given.
So the real change is not that lawyers become obsolete. It is that the purely repetitive layer of legal work becomes less valuable, while legal judgment becomes more valuable.
The Real Risk Is Unverified AI
The biggest danger is not that AI exists. The danger is treating AI output as if it is final.
Anyone using AI for legal work needs to understand hallucinations. AI tools can produce confident, fluent, and completely fabricated information, including case citations that do not exist. That risk is especially serious in law because one fake citation is not just a technical mistake. It can mislead a court, damage a client's case, and create professional consequences for the person who relied on it.
That is why the right mindset is not 'AI did the work for me.' The right mindset is 'AI gave me a draft, and now I have to do the lawyer's work.'
The Supreme Court Reminder: AI Needs a Human Backstop
The stakes became very real in a recent Indian case involving AI-generated precedents. In Gummadi Usha Rani v. Sure Mallikarjuna Rao, the Supreme Court of India took note of a trial court relying on non-existing, fake, or synthetic AI-generated judgments while deciding a property-related dispute.
A bench of Justices P.S. Narasimha and Alok Aradhe treated the issue as a serious institutional concern. The Court issued notice to the Attorney General for India, the Solicitor General of India, and the Bar Council of India to examine the consequences and accountability of relying on AI-generated fake precedents. It also observed that a decision based on non-existent judgments is not merely an ordinary decision-making error; it may amount to misconduct with legal consequences.
That is exactly the point. AI can help legal professionals move faster, but speed without verification is dangerous. In law, the final safeguard cannot be the tool. It has to be the human being using it.
The same concern has also been expressed outside that case. Justice B.R. Gavai has cautioned that AI should assist the judiciary rather than substitute human judgment, because legal disputes require ethical reasoning, empathy, context, and accountability. Those are not optional features of justice. They are the core of it.
What This Means for Young Lawyers and Law Students
So where does that leave a young lawyer or law student who is wondering whether the profession is still worth entering?
My honest answer is this: do not wait for a firm, chambers, or senior to train you on AI. Learn to use it responsibly now. Treat it as a legal skill, not a tech shortcut.
A few practical starting points have helped me:
· Prompt like a lawyer. Give the tool the jurisdiction, facts, issue, procedural stage, desired format, and limits of the task. Vague prompts produce vague drafts.
· Verify like you are signing it. Check every citation, statutory reference, quotation, and factual claim before using it in real work.
· Edit for strategy, not just grammar. AI can produce a clean paragraph, but it cannot know your client's risk appetite, the judge's preferences, or the negotiation dynamics.
· Use saved time to build human skills. Argumentation, negotiation, client trust, courtroom presence, and ethical judgment will matter more, not less.
The lawyers who struggle in the next few years will not necessarily be the ones whose tasks were automated. They will be the ones who either ignored these tools completely or trusted them blindly without developing the judgment to catch their mistakes.
Where Lexi Should Sit in the Workflow
This is also how I think legal AI tools should be positioned inside actual legal work. Lexi should not be treated as the lawyer in the room. It should be treated as the assistant that helps the lawyer reach the important work faster.
For example, a lawyer can use Lexi to get a first draft, identify potential issues in a contract, organize research, or surface relevant authorities. But the lawyer still decides what matters, what is risky, what is persuasive, and what can ethically be relied on.
In other words, AI belongs in the workflow, not in the driver's seat. Used properly, it gives lawyers more time for the part of lawyering that cannot be automated: judgment.
The Bottom Line
AI is not coming for the legal profession in the way people often imagine. It is coming for the repetitive, first-pass work that has traditionally consumed a huge amount of legal time: summarizing, scanning, drafting, organizing, and searching.
That does not make lawyers obsolete. It changes what lawyers must be good at.
The future will not belong to the lawyer who insists on doing everything manually. It will also not belong to a machine working alone. It will belong to lawyers who know how to direct technology, verify its output, and use the time saved to make better legal decisions.
Last summer, I watched a senior associate clear a stack of contracts before lunch and wondered if I had chosen the wrong career. Today, I know I did not. I just need to become the kind of lawyer who can use the tool without surrendering the judgment that makes someone a lawyer in the first place.