AI Sanctions Are Hitting Lawyers in 2026: A Michigan Competence Guide
The tools have gotten good enough that a busy lawyer can produce a polished brief in minutes. The problem is that a confident brief and a correct brief are not the same thing, and courts have stopped treating the difference as an honest mistake. In March 2026 the Sixth Circuit, the federal appellate court that covers Michigan, fined two attorneys $15,000 each for a brief riddled with fabricated citations. This is a plain-English look at what the current rules require of Michigan lawyers who use artificial intelligence, and how to use it without ending up in a sanctions order.
The Sanctions Wave Is Real, and It Reaches Michigan
What started as a handful of embarrassing headlines has become a documented pattern. A public database maintained by researcher Damien Charlotin has logged more than 1,000 United States court decisions addressing AI-fabricated content in filings as of mid-2026, and the pace is climbing, not slowing. See the AI Hallucination Cases database. The penalties have grown with the count.
The case that matters most to Michigan practitioners is Whiting v. City of Athens, No. 24-5918 (6th Cir. 2026). The Sixth Circuit found that counsel had submitted a brief containing more than two dozen fake citations, quotations that did not appear in the cited opinions, and cases cited for propositions they did not support. The court ordered each attorney to pay a $15,000 punitive fine to the court registry, made them jointly responsible for the appellees’ attorney fees, and awarded double costs. The court did not need to prove the lawyers used a chatbot. Its point was broader and it should be printed and taped to every monitor in a law office: no filing should contain a citation, however it was generated, that the signing lawyer has not personally read and confirmed. See the summary at LawSites.
Michigan lawyers who practice only in state court are not insulated. The same conduct has drawn far heavier consequences elsewhere. A federal court in Oregon imposed a $110,000 sanction, the largest AI-hallucination penalty recorded to date, after a filing contained 23 fabricated citations and eight invented quotations. And the Nebraska Supreme Court suspended an Omaha attorney in April 2026 after a divorce appeal in which 57 of 63 citations were defective, including 20 hallucinated cases. The trend line is clear: fines first, then fee-shifting, then license discipline.
What Michigan’s Rules Actually Require
None of this is a new body of law. The obligations that AI misuse violates were already on the books, and Michigan adopted the key one years before generative AI reached the mainstream.
Michigan Rule of Professional Conduct 1.1 requires competent representation. In 2019 the Michigan Supreme Court added a comment making the technology point explicit: a lawyer should keep abreast of the benefits and risks of relevant technology, including “existing and developing technology,” to the extent reasonably necessary for the representation. The State Bar of Michigan has since published an Artificial Intelligence FAQ and a June 2025 Age of AI report confirming that competence now includes understanding what these tools do well, where they fail, and the duty to independently verify their output.
Three other rules do the heavy lifting when a filing goes wrong. MRPC 3.3 imposes a duty of candor toward the tribunal, and a citation to a case that does not exist is a false statement of law to the court. MRPC 5.1 and 5.3 make managing and supervising lawyers responsible for the work of the associates, paralegals, and contract staff they oversee, which means a partner cannot escape a bad filing by pointing at a junior who ran the search. And Michigan Court Rule 1.109(E) is the state analog to the federal signature rule: signing a document certifies that it is well grounded in fact and warranted by existing law, and a violation carries mandatory sanctions. See MCR 1.109. That signature is a personal certification. It does not become someone else’s problem because a machine drafted the first version.
ABA Formal Opinion 512 Is the National Playbook
On July 29, 2024, the American Bar Association issued Formal Opinion 512, its first ethics guidance on generative AI. Michigan has not issued a formal opinion of its own, so 512 is the most useful roadmap available, and it reads onto Michigan’s parallel rules cleanly. Its core message is that using AI is permitted, but every existing duty travels with the technology. The opinion also lands a point that surprises many lawyers on the billing side: a lawyer may bill only for time actually worked, not for the time a tool “saved,” and generally may not bill for the hours spent learning a tool for general use. The work you actually perform, including verifying the output, remains billable.
| Professional duty | What it requires with AI | How AI misuse violates it |
|---|---|---|
| Competence (MRPC 1.1) | Understand the tool’s limits, including hallucination risk, and independently verify output | Filing AI-generated text without knowing the tool invents citations |
| Candor to the tribunal (MRPC 3.3) | Confirm every legal authority is real and stands for the proposition cited | Citing a fabricated case or a real case that says the opposite |
| Confidentiality (MRPC 1.6) | Get informed client consent before entering confidential facts into a tool that may store them | Pasting a client’s file into a public chatbot that trains on inputs |
| Supervision (MRPC 5.1 and 5.3) | Adopt a firm AI policy and check the work of anyone using these tools | Letting staff file AI drafts with no review process in place |
| Reasonable fees (MRPC 1.5) | Bill for work performed, not for time the tool saved | Charging a full research hour for a task the tool did in minutes |
| Certification (MCR 1.109(E)) | Sign only filings you have personally confirmed are grounded in fact and law | Signing a brief you did not read cover to cover |
A Verification Workflow That Keeps You Out of Trouble
The lawyers who get sanctioned are rarely reckless people. They are usually overloaded people who trusted an output that looked authoritative. A short, boring, repeatable process is the entire defense. Building that habit into how the firm trains its people is itself part of the competence and supervision duties, not an optional add-on.
- Treat AI as a first-year associate, not an oracle. It can draft and summarize. It cannot be the final word on whether a case is real or says what the draft claims.
- Pull every citation independently. Open each case in Westlaw, Lexis, or a free primary source and read the passage the draft relies on. If you cannot find it, it is not real, no matter how confident the draft sounds.
- Confirm the case still stands. A real citation can still be wrong if the decision was reversed, vacated, or superseded. Check the treatment, not just the existence.
- Keep client confidences out of public tools. Use a closed or firm-controlled system, or strip identifying facts, unless the client has given informed consent to a specific tool.
- Write it down. A one-page firm AI policy plus a short verification checklist satisfies the supervision duty and gives you something to point to if a filing is ever questioned.
The Stakes Go Beyond the Fine
A sanctions order is a public document. It follows a lawyer into every future case, every malpractice-carrier renewal, and every internet search a prospective client runs. Opposing counsel will cite it. A malpractice insurer that has started asking about AI use on its applications will read it as a claim waiting to happen. And the underlying client, whose brief was built on cases that do not exist, has a real argument that the representation fell below the standard of care. The fine is the cheapest part of the consequence. The reputational and disciplinary tail is what costs a practice.
ਅਕਸਰ ਪੁੱਛੇ ਜਾਂਦੇ ਸਵਾਲ
Can lawyers use AI to write legal briefs?
Yes. No rule and no court prohibits using AI to draft or research. What the rules require is that the lawyer independently verify every citation and factual claim before signing and filing. The problem in the sanctions cases was never the use of a tool. It was filing the output without checking it.
What happens to a lawyer who files AI-hallucinated citations?
Consequences in 2026 have ranged from monetary fines to fee-shifting to bar suspension. The Sixth Circuit fined two attorneys $15,000 each in Whiting, an Oregon federal court imposed a $110,000 sanction, and the Nebraska Supreme Court suspended an attorney’s license. Michigan lawyers face the same exposure under MCR 1.109(E) and the Rules of Professional Conduct.
Does Michigan have a rule requiring lawyers to be competent with AI?
Yes, indirectly. MRPC 1.1 requires competent representation, and a 2019 comment adopted by the Michigan Supreme Court makes clear that competence includes keeping current on the benefits and risks of relevant technology. The State Bar of Michigan has published AI guidance confirming that this now covers generative AI.
Is it a confidentiality violation to put client information into ChatGPT?
It can be. ABA Formal Opinion 512 advises lawyers to obtain informed client consent before entering confidential information into a tool that may store or train on inputs. The safer practice is to use a closed, firm-controlled system or to remove identifying details entirely.
Can I bill a client for the time AI saved me?
Generally no. Opinion 512 concludes that a lawyer may not bill for time the tool saved and generally may not charge a client for hours spent learning a tool for general use. You bill for the work you actually performed, including the verification.
Have a Michigan injury case that outgrew a do-it-yourself approach?
Attorney Manny Chahal handles serious personal-injury and no-fault matters across Michigan and accepts referrals and co-counsel arrangements. If a case needs a partner who will do the work and verify it, start a conversation.
1-844-624-2425 'ਤੇ ਕਾਲ ਕਰੋ