Can Lawyers Use AI? A Country-by-Country Tracker (2026)
Yes, in most major markets, with duties attached. We tracked 17 jurisdictions: 10 permit with guidance, 1 restricts (Qatar), 6 say nothing at all.
The Short Answer
Can a lawyer use AI? In almost every major legal market, yes, provided they treat the output as a draft, not an oracle. The recurring theme across every regulator that has spoken is the same: AI does not dilute a lawyer's existing duties. You still owe the client competence, confidentiality, and a duty to verify. The tool changes; the responsibility does not.
Key facts
- Across 17 jurisdictions tracked: 10 permit lawyer AI use with formal guidance, 1 imposes court-enforced verification duties (Qatar), 6 have not addressed it.
- Qatar's QICDRC Practice Direction No. 1 of 2026 requires every citation to be checked against primary sources, but it does not require lawyers to declare AI use unless the court asks.
- In England and Wales, the SRA's warning notice of 17 August 2026 says AI "does not diminish or transfer your professional responsibilities", and the Law Society's October 2026 practice note says solicitors must keep the ability to do and check the work without AI.
The American Bar Association set the template with Formal Opinion 512 in 2024: understand the tool's limits, protect client confidences, verify the output, and bill reasonably. Most other regulators that followed echo it. The interesting story is not the consensus. It is the map of where regulators have said nothing at all.
The Tracker: 17 Jurisdictions
We ran live searches across 17 jurisdictions and classified each as permitted with guidance, restricted, or not yet formally addressed. Every classification traces to a regulator instrument or its documented absence.
| Jurisdiction | Status | Instrument |
|---|---|---|
| United States | Permitted | ABA Formal Opinion 512 (2024) + state bar opinions |
| United Kingdom | Permitted | Law Society practice note (Oct 2026), SRA warning notice (Aug 2026), Bar Council guidance |
| European Union | Permitted | No bar rule; AI Act high-risk covers court use, not firm tools (2 Dec 2027) |
| France | Permitted | CNB déontologie guide (Mar 2026) |
| Germany | Permitted | BRAK Leitfaden on AI in law firms (Dec 2024) |
| Canada | Permitted | Law Societies of AB / ON / BC (2024–26) |
| Australia | Permitted | Law societies + SA & Federal Court rules (2026) |
| Brazil | Permitted | OAB Rec. 001/2024 + CNJ Res. 615/2025 |
| Singapore | Permitted | MinLaw GenAI legal-sector guide (Mar 2026) |
| India | Unaddressed | No BCI rule; Supreme Court notice (Feb 2026) |
| Qatar | Restricted | QICDRC Practice Direction 1/2026, citation verification |
| United Arab Emirates | Permitted | DIFC Courts PGN 2/2023; ADGM: Arabyads v Alam costs order (Dec 2025); onshore courts: no AI rule |
| Saudi Arabia | Unaddressed | SDAIA guidelines only; no bar rule |
| Egypt | Unaddressed | Bar training event (Dec 2025); no rule |
| Lebanon | Unaddressed | Beirut Bar AI committee + 2026 MoU; no rule |
| Jordan | Unaddressed | Tech & data laws only; no bar rule |
| Morocco | Unaddressed | No bar or court guidance found |
The MENA Picture
Is it ethical for lawyers to use ChatGPT? The duty of competence
The question lawyers actually type into a search bar is narrower than 'can lawyers use AI.' It is 'is it legal for lawyers to use ChatGPT,' and the honest answer is: yes, but the bar moved the goalposts. In the United States, using a general-purpose chatbot is not prohibited. What is prohibited is using one badly. The American Bar Association did not write a new rule for AI; it pointed the existing Model Rules of Professional Conduct at it. ABA Formal Opinion 512, issued July 29, 2024, runs the analysis through Rule 1.1 (competence), Rule 1.6 (confidentiality), Rule 5.3 (supervision of non-lawyer assistance), Rule 3.3 (candor to the tribunal), and the reasonable-fee rules.
Competence is the one most lawyers underestimate. Opinion 512 says a lawyer must understand a tool's 'benefits and risks' before using it, and it adds a line most people skip: this is not a one-time box to tick. In the ABA's words, 'technological competence presupposes that lawyers remain vigilant about the tools' benefits and risks,' because the tools change underneath you. A lawyer who cannot explain, in plain terms, that a consumer chatbot generates plausible text rather than retrieving verified law is not yet competent to use it on a client matter. That is not our opinion. That is the framing the regulator chose.
By March 2026, more than 35 US state bar associations had issued their own AI guidance, and they echo the same spine: understand the tool, protect the client, verify the output, bill honestly. The pattern is identical to what every regulator in our tracker above arrived at. The tool is allowed. The duty is non-delegable. You can hand the typing to a model; you cannot hand it the responsibility.
Two parts of Opinion 512 fall on the firm rather than the individual lawyer. Under Rules 5.1 and 5.3, managerial lawyers are expected to set clear policies on which AI tools are approved and for what, and supervisory lawyers must make sure everyone using them, non-lawyer staff and outside vendors included, is trained to use them and to check the output. On fees, the opinion says a lawyer billing hourly may charge only for the time actually spent, including time spent reviewing AI output, and may not bill for the hours the tool saved; general learning time on a tool the lawyer will use across clients is overhead, not a client charge. Outside the bar rules, the frameworks firms cite most for their own AI policies are UNESCO's Recommendation on the Ethics of Artificial Intelligence (2021) and the EU High-Level Expert Group's Ethics Guidelines for Trustworthy AI (2019), both of which put human oversight and accountability at the centre.
The confidentiality risk: consumer ChatGPT vs a legal-grade tool
This is where 'can I use ChatGPT' quietly becomes 'should I.' The single hardest duty for a chatbot to satisfy is Rule 1.6 confidentiality. Opinion 512 is direct about it: because many self-learning AI tools are 'designed so that their output could lead directly or indirectly to the disclosure of information relating to the representation of a client,' a lawyer must get the client's informed consent before inputting that information into such a tool. Not a line buried in an engagement letter. The ABA specifically flags that boilerplate consent 'may not suffice.'
The reason is architectural, not legal. On consumer ChatGPT (Free and Plus), the default is that OpenAI may retain your prompts and use them to train its models unless you turn that off, and conversations are typically retained for a period regardless of the training toggle. The same is true of consumer-tier Claude and Gemini. Opting out of training does not erase the provider's right to retain data or disclose it in response to legal process. The moment client material lands on that infrastructure, you have arguably shared a privileged communication with a third party who owes your client nothing.
That risk is not theoretical. In United States v. Heppner (S.D.N.Y., February 2026), a federal court held that documents a litigant generated with a consumer-tier AI tool and later shared with counsel were neither privileged nor protected as work product: an AI tool is not an attorney, holds no license, and owes no duty of confidence. We unpack that decision in our piece on why AI conversations are not privileged.
A legal-grade tool closes the gap by contract and by construction. Enterprise and commercial tiers (ChatGPT Enterprise, Claude's commercial plans) exclude your inputs from training by default and attach contractual confidentiality. Purpose-built legal platforms go further: zero-data-retention agreements with the model providers, privileged material kept under the firm's control, and an audit trail of what the AI did and what a human approved. The distinction that matters is not 'AI vs no AI.' It is 'a tool you can give a confidentiality guarantee on vs a free consumer product where, when the tool is free, your client's data is the product.'
We tested what this looks like in practice: the same NDA review run on ChatGPT and a purpose-built legal AI, and where generic AI falls short of legal-specific tooling.
The Gulf is further along than the headlines suggest, and not in the order most summaries claim. Qatar's QICDRC Practice Direction No. 1 of 2026, issued on 6 January 2026, is the region's most-cited instrument, but the Dubai International Financial Centre Courts got there two years earlier, with a guidance note in force since December 2023. On disclosure the two run opposite to expectation: the DIFC asks parties to declare AI use at the earliest opportunity, while Qatar's paragraph 24 says pre-emptive declaration is not required unless the court specifically asks for it. What Qatar does impose without condition is verification: every citation checked against primary sources or a recognised legal database. It emerged from an actual case, not a think-tank.
Elsewhere the picture is uneven. Abu Dhabi's ADGM has no AI practice direction, but in December 2025, in Arabyads Holding Ltd v Gulrez Alam [2025] ADGMCFI 0032, its Court of First Instance ordered the defendant's former legal representatives to pay AED 282,508 in wasted costs, on an indemnity basis, for filing unverified AI-assisted research: enforcement under existing conduct rules rather than a new instrument. Onshore federal and local courts had issued no AI guidance for practitioners as of March 2026 (Oxford Institute of Technology and Justice), so a Dubai lawyer can work under three regimes in one week. We map them in Onshore, DIFC, ADGM. Saudi Arabia has SDAIA guidelines aimed at AI governance generally; we read the Rules of Professional Conduct for Lawyers and their 2024 amendment and found no AI provision in either. The Beirut Bar Association has an AI committee and a 2026 memorandum of understanding with the Lebanese government. The Egyptian Bar ran a 'Generative AI for Lawyers' training event in late 2025. For Morocco and Jordan we found awareness but nothing directed at lawyers.
For a region HAQQ is built for, that is the whole opportunity: in markets where the norm is not yet codified, the firms that adopt AI responsibly now will set the standard others get measured against.
What It Means
'Is it allowed?' is the wrong question to stop on. Every regulator that has spoken says the same thing in different words: you may use AI, and you remain fully responsible for the result. So the real question is whether your tools make that responsibility easy to discharge: citations you can check, confidentiality you can guarantee, a record of what the AI did and what a human approved.
That is the bet we made with HAQQ: build for the duty, not around it. The Gulf regulators have already started writing these rules, and every instrument that exists so far points the same way: verify the output, and keep the responsibility. Tools built around verification are aligned with them by construction. The rest will be scrambling.
Key Takeaways
- Across the 17 jurisdictions tracked here: 10 permit with guidance, 1 restricts, 6 unaddressed.
- Every regulator agrees AI does not reduce a lawyer's duties.
- The DIFC Courts moved first (2023); Qatar wrote the region's first dedicated AI practice direction (2026).
- Early responsible adoption sets the regional standard.
Sources & Further Reading
- 2,046 court cases with AI-fabricated citations
- the sanctions record: 496 attorneys, $55K in fines
- our jargon-free LLM guide for lawyers
- ABA Formal Opinion 512: generative AI and ethics rules
- EU AI Act: regulatory framework
- Qatar QICDRC: Practice Direction on the Use of Artificial Intelligence (No. 1 of 2026)
- DIFC Courts: Practical Guidance Note No. 2 of 2023 on LLMs and generative AI
- BRAK (Germany): AI guidance for law firms
- SRA: Warning notice, misuse of AI (17 August 2026)
- Legal Cheek: AI is no substitute for junior lawyers, Law Society warns (22 October 2026)
- Oxford Institute of Technology and Justice: UAE AI Justice Atlas (March 2026)
- Morgan Lewis: ADGM Courts judgment serves as a warning regarding proper use of AI (February 2026)
FAQ
Can lawyers use AI like ChatGPT for legal work?
In most major jurisdictions, yes, under existing duties of competence, confidentiality, and verification. Several regulators, such as the ABA's Formal Opinion 512, have issued explicit guidance.
Do lawyers have to disclose AI use to the court?
It depends on the jurisdiction, and the common assumption is often backwards. Qatar's QICDRC practice direction expressly does not require pre-emptive declaration of AI use unless the court asks for it, though disclosure is unconditional for AI-generated or enhanced audio and visual evidence. The DIFC Courts go further and ask parties to declare AI use at the earliest opportunity. Most courts ask for verification rather than blanket disclosure, and most jurisdictions still have no rule at all.
Is it legal for lawyers in the UAE or Saudi Arabia to use AI?
Yes. No bar-association rule in either country prohibits it as of September 2026. But the UAE is not a blank slate: the DIFC Courts have had guidance on large language models and generative AI in proceedings since December 2023, and in December 2025 the ADGM Court of First Instance ordered a firm to pay AED 282,508 in indemnity costs for filing unverified AI-generated authorities. Saudi Arabia's Rules of Professional Conduct for Lawyers contain no AI provision, so general professional duties apply.
Is it ethical for lawyers to use ChatGPT?
Yes, where you satisfy the existing rules of professional conduct. ABA Formal Opinion 512 (July 2024) runs AI use through the duties of competence (Rule 1.1), confidentiality (Rule 1.6), supervision (Rule 5.3), and candor to the court (Rule 3.3). It is ethical to use ChatGPT as a drafting aid if you understand its limits, protect client confidences, verify every output, and bill reasonably. It becomes unethical the moment you skip any of those, most commonly by failing to verify a citation or by inputting confidential client data into a consumer tool.
Is it legal for lawyers to use ChatGPT?
In most major jurisdictions, yes. No US bar association prohibits AI use; instead, by March 2026 more than 35 state bars had issued guidance applying the existing Model Rules to it. Qatar's QICDRC practice direction is the notable exception in our tracker, requiring every citation to be checked against primary sources. The EU AI Act is often cited here but its high-risk rules for the justice sector target AI used by or on behalf of a judicial authority, not a law firm's own research tool, and they do not apply until 2 December 2027. 'Legal' is the easy part: the duty to verify the output and protect client data travels with you regardless of jurisdiction.
Does ABA Formal Opinion 512 require client consent to use AI?
For confidential information, effectively yes. Opinion 512 states that because many self-learning AI tools are designed so their output could lead to disclosure of client information, a lawyer must obtain the client's informed consent before inputting that information into such a tool. It also warns that boilerplate consent in an engagement letter may not suffice. There is no blanket duty to disclose AI use for non-confidential drafting, but court standing orders may impose one.
Can I put client information into ChatGPT without breaching confidentiality?
Not on the consumer tier without real risk. Free and Plus ChatGPT may retain prompts and use them to train OpenAI's models unless you disable training, and even then the provider retains data and can disclose it under legal process. Inputting privileged client material there arguably shares it with a third party that owes your client no duty of confidence. If you must use a general tool, strip identifying details first; for real client work, use an enterprise or purpose-built legal tool with a contractual confidentiality or zero-data-retention guarantee.
What is the difference between consumer ChatGPT and a legal-grade AI tool for confidentiality?
Consumer ChatGPT trains on your inputs by default and offers no contractual confidentiality, so client data flows to a third party with no duty to protect it. Legal-grade tools (enterprise tiers or purpose-built legal platforms) exclude inputs from training by default, attach contractual confidentiality, often run under zero-data-retention agreements, and keep an audit trail of what the AI did and what a human approved. The difference is whether you can give your client an honest confidentiality guarantee.
Do lawyers have to verify ChatGPT's output before filing it?
Yes, and this is the duty courts enforce hardest. The recurring lesson from AI sanctions cases is that the problem was never that the lawyer used AI: it was that the lawyer failed to verify the output, which courts treat as a breach of candor and of reasonable inquiry, not a technology glitch. Confirm every cited authority against a primary source before filing. The duty to verify is non-delegable and sits at signature, not just at drafting.
What is ABA Formal Opinion 512?
ABA Formal Opinion 512 is the American Bar Association's ethics opinion on lawyers' use of generative AI, issued on 29 July 2024. It wrote no new rule. It applies the existing Model Rules: competence (Rule 1.1), confidentiality (Rule 1.6), supervision (Rule 5.3), candor to the tribunal (Rule 3.3) and the reasonable-fee rules. It also treats technological competence as ongoing, because the tools change underneath the lawyer.
Can lawyers bill for time saved by AI under ABA Formal Opinion 512?
No. Under ABA Formal Opinion 512 (July 2024), a lawyer billing hourly may charge only for the time actually spent, including time spent reviewing AI output, and may not bill for the hours the tool saved. General learning time on a tool the lawyer will use across clients is overhead, not a client charge.
What does ABA Formal Opinion 512 require of law firms?
Clear policies and training. Under Rules 5.1 and 5.3, ABA Formal Opinion 512 (July 2024) expects managerial lawyers to set clear policies on which AI tools are approved and for what. Supervisory lawyers must make sure everyone using them, non-lawyer staff and outside vendors included, is trained to use the tools and to check the output.