- within Litigation, Mediation & Arbitration, Tax and Accounting and Audit topic(s)
- with Finance and Tax Executives
- in Asia
- with readers working within the Accounting & Consultancy and Retail & Leisure industries
Historically, a young associate’s edge came from being the hardest worker in the room. The first one in, the last one out, the one who reads every footnote and knows the facts of the case by heart. That hasn’t changed, but today there’s an added layer: artificial intelligence. How and when you use AI tools will say a lot about the kind of lawyer you’re becoming.
The good news is that AI, used prudently, is one of the most powerful tools a young attorney has ever had access to. The not-so-good news is that used carelessly, it can be professionally damaging and even ethically perilous. Here are some tips for using AI responsibly and effectively.
Embrace AI, But Know the Rules
I recently spoke with several law students about this emerging technology and heard a range of reactions from hesitation to staunch avoidance. Some expressed concerns that offloading cognitive function to these tools would diminish the development of their own legal judgement. Others offered the sentiment that leaning on AI signals laziness. While I understand the instinct, I counter that these views are both mistaken and potentially costly.
AI isn’t going anywhere. It’s becoming more capable by the month. Attorneys who learn to use it effectively will have an edge over those who don’t.
With that said, navigating the ethical and responsible use of AI begins with a few non- negotiable rules:
- Always verify. Treat AI like an enthusiastic but inexperienced intern. It can synthesize information quickly and draft competently, but it makes mistakes, including confidently citing cases that don’t exist, misquoting holdings, or getting a statute slightly wrong. Every cite, source, quote, and legal proposition generated by AI must be independently verified before it goes anywhere with your name on it. An unverified AI hallucination in a brief is not just embarrassing; it’s an ethical violation waiting to happen.
- Know your firm’s policy. Before using any AI tool for client work, confirm that your firm or organization permits it. Policies vary widely, and ignorance is not a defense.
- Never upload confidential information to an open-universe AI. If you’re using a open-facing AI tool, assume anything you type and upload could be retained, reviewed, or exposed. If your firm uses a closed, enterprise AI system, confirm the security protocols before inputting any client data.
- Disclose and document. Be transparent with your supervising attorney about when and how you used AI, and confirm that you’ve verified the accuracy of anything AI-generated. This builds trust and protects you both. See also ABA Formal Op. 512 on AI and client disclosure, as well as the growing number of federal/state court judges with standing orders requiring disclosure of AI use in filings.
The Skill Nobody Taught You in Law School: Writing Good Prompts
If there’s one AI skill that will separate effective legal practitioners from ineffective ones in the years ahead, it’s the ability to write a precise, well-structured prompt. Good prompt writing, at its core, is simply good communication. The discipline required to write a precise, well-structured prompt is the same discipline required to write an email to a colleague, a persuasive argument to a judge, or a plain- language explanation to a client. If you find yourself struggling to articulate what you need from AI, that’s often a signal that you haven’t yet fully clarified your own thinking. In that sense, the prompt is a useful diagnostic: if you can’t explain it to AI, you may not be ready to explain it to anyone else either. Prompt writing is a skill that can be learned like any other, and small improvements in how you frame a
request can produce dramatically better results. The key principles are specificity, context, and structure: tell AI what you need, why you need it, what format you want the output in, and what constraints apply. Consider the difference between these three approaches to the same task: Weak prompt: “Helpmeprepareforadeposition.” Medium prompt: “Draft deposition questions for the CFO about the timeline of the attached contract.”
Strong prompt: “I am deposing the defendant’s CFO in a breach of contract dispute. The central issue is whether the defendant knew, at the time of signing, that it could not meet the delivery timeline specified in the contract. Based on the attached documents, which include the signed contract, internal emails from the defendant’s finance team, and the CFO’s prior deposition in an unrelated matter, please act as an experienced litigator preparing for a high-stakes deposition and draft a deposition outline that: (1) establishes the CFO’s role and responsibilities at the time of contracting; (2) builds a foundation for the internal emails; (3) pins the witness to key representations about the timeline; and (4) identifies gaps or ambiguities in the documents that the deposition should attempt to fill, and suggests questions designed to address them.”
The quality of your AI output is a direct function of the quality of your input, and learning to write prompts with the same precision you bring to a well-drafted interrogatory is itself a professional skill worth developing deliberately.
Beyond Drafting: Unexpected Ways AI Can
Work for You
AI can be used for more than just drafting and research. It is remarkably effective at building chronologies and timelines from large volumes of documents. Feed it a set of emails, deposition excerpts, or medical records and ask it to extract and sequence every date-referenced event, and you’ll have a working timeline in minutes rather than hours. (But you still need to take time to verify the accuracy!)
It can summarize lengthy contracts, deposition transcripts, or expert reports into structured, issue-specific digests, letting you orient yourself quickly before diving into the details. AI can also help you stress-test your own arguments by playing the role of opposing counsel or skeptical judge. Prompt it to write the strongest possible response to your motion and you may find vulnerabilities you hadn’t considered.
You can even use AI to analyze deposition transcripts for internal inconsistencies, flagging places where a witness’s testimony diverges from prior statements. These use cases can accelerate discovery and research, but AI does not replace your legal judgment. It can, however, free up some time and mental bandwidth to enable more impactful work.
One final tip worth emphasizing: AI is a remarkably effective proofreading tool, and not just for grammar and spelling. A well- constructed prompt can direct it to catch internal inconsistencies, undefined or inconsistently used defined terms, mismatched cross- references, and formatting irregularities that are easy to overlook when you’ve been staring at the same document for hours. However, as always, be mindful of confidential information and make sure you are using workplace approved AI tools only.
AI Can’t Replace You. Don’t Let It Try.
For all its capabilities, AI has no judgment. AI doesn’t know your client. It can’t read the room in a deposition, sense when opposing counsel is bluffing, or understand why a particular client is risk-averse in a way that changes your strategic approach. It has no empathy, no intuition, and no stake in the outcome.
Those are your advantages. Don’t give them up.
One of the easiest traps is over-delegating to AI, particularly for things like client emails, demand letters, and other correspondence. The temptation is understandable: you’re busy, AI drafts quickly, and the output is usually serviceable. But serviceability shouldn’t be the goal. When you outsource all your thinking to AI, you stop developing the muscles you’ll need for the rest of your career.
Here’s a better approach: write the first draft yourself. Then ask AI to push back on it. You might prompt something like: “Act as opposing counsel and review this demand letter. Draft sample response pointing out weak logic, unsupported claims, or emotional appeals. Do not suggest edits to my draft; allow me to redraft based on your output.” Let AI play devil’s advocate while you do the thinking.
The same principle applies to legal arguments, memos, and briefs. Your clients pay for your voice, your analysis, and that is what partners are evaluating you on as well. When used well, AI will help you sharpen your own thinking, not replace it.
The Bottom Line
The attorneys who will thrive in the years ahead are those who bring the same analytical rigor to their use of technology as they do to their legal work. The attorneys who get this right will use technology to amplify their best work. Those who get it wrong will use it as a substitute for their own thinking. The difference between the two is the same difference that has always separated good lawyers from great ones.
The fundamentals haven’t changed: know the law, know your client, show up, do the work, and earn trust. AI is a powerful new tool that can act in service of those fundamentals, but it is not a replacement for who you are.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.