ChatGPT for Lawyers: 10 Use Cases, Sample Prompts, and the Risks to Manage

ChatGPT for lawyers is a productivity tool for first drafts, document summaries, client emails, and exploratory legal research, not a substitute for legal judgment. Lawyers can use it ethically under ABA Formal Opinion 512 if they protect client confidentiality, verify every citation, and remain responsible for the final work product.

This guide covers ten use cases ranked from quick wins to advanced work, five short prompts, the four risks behind real court sanctions, what the ABA and state bars require, and a compliance checklist you can adopt this week.

Can lawyers use ChatGPT? Where US law firms stand

Yes. No state bar prohibits it, and ChatGPT is the tool most lawyers start with, especially in smaller practices.

The adoption numbers

The ABA's 2024 Legal Technology Survey Report found that 30% of responding lawyers used AI-based tools, up from 11% in 2023. Among firms using or considering AI, ChatGPT led at 52%, according to a LawSites summary of the survey.

Small practices lean on it most: 62% of solo practitioners and 64% of firms with 2 to 9 attorneys were using or considering ChatGPT, against 36% of firms with 100 or more attorneys. Accuracy was the top concern, cited by 75% of respondents.

Your existing ethics rules already apply

ABA Formal Opinion 512 (July 2024) confirms that you may use generative AI if you meet your usual duties, from competence to reasonable fees. Our guide to AI for lawyers covers the broader landscape, and our article on ChatGPT in legal practice: risks and best practices lays the groundwork.

10 ChatGPT use cases for law firms

Here are ten realistic uses, from the most immediate to the most advanced. One rule applies to all of them: no identifiable client information on a consumer account. Anonymize first, or use a business plan your firm has vetted.

Drafting and writing

• 1. First drafts of demand letters, motions, and memos. A structured draft that you correct, support with real authority, and sign.
• 2. Client emails. Ask for a clearer version of your draft reply in plain English. You keep the legal position; the tool only rewrites.
• 3. Website FAQs. A strong SEO lever when structured with clear H1, H2, and H3 tags. Our guide to building a law firm blog details the method.
• 4. Social media posts. Turn a recent ruling into a post, within the limits of Rule 7.1 on communications about your services. See our guide to LinkedIn for lawyers.

Reading, summarizing, and research

• 5. Document summaries. Deposition transcripts, discovery responses, or a 60-page lease condensed into facts, dates, and open issues.
• 6. Exploratory legal research. Map an unfamiliar area or generate search terms for Westlaw or Lexis+, never as a source of citations.
• 7. Clause-by-clause contract review. Spot unusual terms and missing protections in an anonymized agreement. Dedicated tools go further, as our guide to AI contract review explains.

Preparation and firm operations

• 8. Meeting notes and voice memos. A dictated memo turned into a clean summary with next steps and deadlines.
• 9. Intake scripts. The questions your staff asks prospective clients on a first call. More in our guide to client intake automation.
• 10. Hearing and deposition prep. ChatGPT plays opposing counsel and points out the weak spots in your argument.

Is ChatGPT good for legal research?

It is good at the first hour of research and unreliable at the last mile.

Where it helps

ChatGPT explains an unfamiliar doctrine in plain English, lists the elements of a claim, suggests counterarguments, and turns a fact pattern into search queries.

Where it fails

It has no access to Westlaw or Lexis+ and no citator. A Stanford RegLab and HAI study found that general-purpose models hallucinated 69% to 88% of the time on verifiable questions about federal cases, and a 2024 follow-up found errors in more than 17% of answers even from Lexis+ AI. Pull each authority, read it, and run it through KeyCite or Shepard's. Our guide to AI legal research compares the dedicated platforms.

5 short ChatGPT prompts for lawyers

A generic prompt gets a generic answer. A useful legal prompt gives a role, a jurisdiction, an expected format, and an instruction to flag uncertainty. It never includes names, case numbers, or other identifying details.

The prompts

• Demand letter. "You are a California employment lawyer. Draft a demand letter to a former employer for unpaid overtime. Leave the facts in [brackets] and write [VERIFY] next to any statute or case you are not certain exists."
• Document summary. "Summarize this anonymized deposition excerpt in 10 lines, separating facts, key dates, and legally relevant admissions. Flag any ambiguity."
• Website FAQ. "You are a Texas family lawyer. Write 8 questions and answers for a law firm website in plain English, with no promotional language and no promise of results."
• Client email. "Rewrite this draft reply as a clear, courteous 150-word message without jargon. Do not add any new legal position."
• Hearing prep. "Based on this anonymized summary, list 10 questions opposing counsel could ask, suggest an answer for each, and point out the weakest parts of my argument."

Why the [VERIFY] flag matters

A generative model can state something false with the same confidence as something true. Asking it to flag uncertainty does not eliminate hallucinations, but it tells you where to look first. For a full library organized by task, see our collection of AI prompts for lawyers.

The 4 biggest risks of ChatGPT for lawyers, and how to neutralize them

Four risks account for almost every problem, and each has a concrete countermeasure.

Hallucinated citations and the sanctions that followed

In Mata v. Avianca (S.D.N.Y., June 2023), Judge P. Kevin Castel fined two New York lawyers and their firm $5,000 for a brief citing cases ChatGPT had invented. The sanction punished the failure to check the output, not the use of the tool.

Sanctions have escalated since. In Wadsworth v. Walmart (D. Wyo., February 2025), a Morgan & Morgan lawyer lost his pro hac vice admission and was fined $3,000, and the two lawyers who signed without reviewing paid $1,000 each. In Johnson v. Dunn (N.D. Ala., July 2025), three Butler Snow lawyers were removed from the case and referred to the Alabama State Bar over unverified ChatGPT citations.

In Noland v. Land of the Free (September 2025), the California Court of Appeal imposed a $10,000 sanction in its first published opinion on AI-fabricated citations. Researcher Damien Charlotin's public database counted 1,058 US cases involving improper AI use in court filings by early June 2026.

Client confidentiality: consumer plans vs. Business and Enterprise

Rule 1.6(c) requires reasonable efforts to prevent the unauthorized disclosure of client information, and with ChatGPT the plan is the first of those efforts. On consumer plans (Free, Plus, and Pro), OpenAI may train on your chats unless you turn off "Improve the model for everyone" in Data Controls. ChatGPT Business (called Team until August 2025), Enterprise, and Edu do not train on your data by default.

Retention matters too. In the New York Times copyright case, a federal court ordered OpenAI to preserve ChatGPT output logs, deleted chats included, from May until September 26, 2025. According to OpenAI, the order covered Free, Plus, Pro, and Team accounts, but not Enterprise or Edu.

Opinion 512 adds that entering client information into a self-learning tool requires informed consent, and boilerplate in an engagement letter is not enough. Removing names, dates, and case numbers remains the simplest safeguard.

Over-reliance and generic answers

Responsibility does not move to the software. Under Rules 5.1 and 5.3, supervising lawyers must make sure everyone who uses ChatGPT follows firm policy, and in federal court a Rule 11 signature certifies that the legal contentions are warranted by existing law. Wadsworth and Johnson v. Dunn show that signing a brief you did not check is sanctionable in itself.

The fourth risk is quieter: without a precise prompt, ChatGPT blends federal and state law into flat, sometimes wrong content. Name the jurisdiction, the court, and the format, then have a lawyer who knows the area review it. To turn outputs into content that attracts clients, see our guide to law firm content marketing.

What ABA Formal Opinion 512 and state bars require

The framework is now clear enough to write a firm policy from. Opinion 512 sets the national baseline, and state bars add their own details.

ABA Formal Opinion 512 applied to ChatGPT

• Competence (Rule 1.1): understand what ChatGPT can and cannot do, hallucinations included.
• Confidentiality (Rule 1.6): evaluate the plan's data terms, and get informed consent before entering client information into a self-learning tool.
• Communication (Rule 1.4): tell the client when they ask, or when AI use matters to a significant decision in the matter.
• Candor and supervision (Rules 3.3, 5.1, and 5.3): verify everything you file, and set policies for the people you supervise.
• Fees (Rule 1.5): bill hourly work only for time actually spent, prompting and review included, and generally not for learning a general-purpose tool.

California, Florida, and New York

The State Bar of California's practical guidance (November 2023) bars entering confidential information into a tool without adequate confidentiality and security protections, asks lawyers to anonymize client details, and rules out hourly fees for time saved by AI.

Florida's Ethics Opinion 24-1 (January 2024) recommends informed consent before sharing confidential information with a third-party AI tool and requires client-facing chatbots to identify themselves as AI. In New York, the State Bar's AI task force report, adopted in April 2024, recommends telling clients when AI tools are used in their matters.

Engagement letters and court orders

Mentioning AI in your engagement letter is good communication, but under Opinion 512 a general clause does not replace specific informed consent. Courts add their own layer: since Judge Brantley Starr's May 2023 standing order in the Northern District of Texas, a number of federal judges require lawyers to certify or disclose AI use in filings.

ChatGPT vs. legal AI tools: which AI tool is best for law?

There is no single best tool, only the right tool for each task: a general-purpose assistant for productivity, a legal-specific platform for case law and client matters.

General-purpose assistants

• ChatGPT (OpenAI): strong for rewriting, structuring, and brainstorming.
• Claude (Anthropic): well regarded for long documents and careful drafting.
• Gemini and Microsoft Copilot: most useful if your firm runs on Google Workspace or Microsoft 365. See Gemini for lawyers and Microsoft Copilot for lawyers.

Legal-specific platforms

CoCounsel from Thomson Reuters, Lexis+ AI, Harvey, Spellbook for contracts, and vLex's Vincent AI are built on verified legal content, with stronger data commitments. They cost more, but they cite sources you can check. Our roundup of legal AI tools compares them.

In short: ChatGPT for non-sensitive tasks, a legal platform as soon as the work involves case law or client files. The same logic shapes how AI assistants describe your firm, as our guide to AI search optimization for law firms explains.

Is AI better than a lawyer?

No. ChatGPT is a drafting and research aid; the lawyer remains the one who advises, decides, and answers for the result.

What ChatGPT cannot do

It cannot take responsibility for advice, weigh strategy against a client's goals, negotiate, or appear in court. OpenAI's usage policies, updated in October 2025, bar tailored legal advice without the appropriate involvement of a licensed professional. And in July 2025, OpenAI CEO Sam Altman acknowledged that ChatGPT conversations carry no legal privilege and could be produced in a lawsuit.

AI shifts legal demand instead of replacing it

The good news for firms: AI moves demand rather than drying it up. Many people who research their problem with ChatGPT realize they have a claim, then look for a lawyer. A well-positioned firm captures them, as we explain in our guides to law firm marketing and getting cited by ChatGPT. For the longer view, read whether AI will replace lawyers.

A compliance checklist for using ChatGPT at your firm

Three moves get you started without risk: write the policy, train the team, and pick the right pair of tools. Training matters most, because the sanction cases above all started with lawyers who trusted output they had not checked.

Before anyone opens ChatGPT

• Choose the plan. Business or Enterprise for anything close to client matters; consumer accounts only for public, non-client tasks.
• Write a one-page AI policy. Approved tools, approved uses, and the data that never goes into a prompt: names, case numbers, health or financial records, privileged communications.
• Do vendor due diligence. Where data is stored, how long it is kept, who can access it, whether it trains models, and whether the vendor holds a SOC 2 report.
• Update engagement letters. Explain AI use plainly, and get specific informed consent where Opinion 512 or your state requires it.
• Train and supervise. Show your team a real hallucination and name a partner who answers questions.

Every time you use it

• Anonymize before you paste, even on a business plan.
• Turn off model training or use Temporary Chat on any consumer account.
• Read and cite-check every authority before it leaves the firm.
• Check the judge's standing order on AI disclosure before filing.
• Bill actual time only, and note which work was AI-assisted.

Then think downstream: once your content is produced, it still has to be found. That is the purpose of a structured visibility strategy, and the work we do every day as a law firm marketing agency.

Key takeaways

ChatGPT saves real time on drafting, summarizing, and preparation, as long as no identifiable client information goes into a consumer account and every authority is verified at the source. For case law and client files, a legal-specific platform or a vetted Business or Enterprise plan is the safer choice. In every case, judgment and responsibility stay with the lawyer.

ChatGPT is neither a threat nor a magic wand: it is a powerful assistant when your expertise, Opinion 512, and your state bar rules frame its use. To turn these uses into visibility and new cases, tell us about your firm.