
Newsroom·Michigan Legislature·News
Disability Claim Survives in University of Michigan AI Misconduct Case
A federal judge let a Michigan student with anxiety and OCD, accused of using AI on her papers, pursue her disability discrimination claim. What it means for you.
A careful way of writing is not a confession.
A federal judge in Michigan ruled on Monday that a University of Michigan undergraduate who was accused of using AI on her course papers may press her claim that she was discriminated against because of her disabilities. In an opinion dated November 9, 2026, U.S. District Judge Laurie J. Michelson of the Eastern District of Michigan dismissed most of the student’s case. She kept alive one claim: intentional disability discrimination under the Americans with Disabilities Act and the Rehabilitation Act.
The student, who sues as Jane Doe, says she lives with “generalized anxiety disorder (GAD) and obsessive-compulsive disorder (OCD).” She says the “tone, meticulous structure, [and] stylistic consistency” of her writing can be misread as machine-made. She was accused of using AI three times in a single course. Monday’s ruling doesn’t say her school was wrong. It says her claim is the kind the law recognizes, and that it deserves to be heard.
01
What Judge Michelson decided on November 9

The ruling is an opinion and order granting in part and denying in part the defendants’ motion to dismiss in Doe v. Regents of the University of Michigan, Case No. 2:26-cv-10451. The defendants are the university’s Regents, the graduate student instructor who brought the accusations, the course’s lead instructor, and two officials in the university’s Office of Student Academic Affairs, known as OSAA. The student accused of using AI sued the four individuals in both their official and individual capacities.
Judge Michelson sorted the claims into three outcomes:
- The procedural due process claim is dismissed. The four individuals get qualified immunity in their individual capacities, and the court found the process the student received was adequate.
- The failure-to-accommodate and retaliation claims under the disability laws are dismissed.
- The claim of intentional disability discrimination under the ADA and the Rehabilitation Act survives against the defendants in their official capacities, along with her request for a declaratory judgment.
The decisive sentence is short: “Plaintiff plausibly pleads a claim for disability discrimination under the ADA/RA.” That isn’t a finding that anyone discriminated. It’s a finding that the complaint of a student accused of using AI, taken as true at this early stage, states a claim the law recognizes. The case now goes forward on that claim, and the opinion sets no trial date or schedule.
The court also denied her request to amend her complaint, without prejudice, because she hadn’t filed a proposed amended complaint. The ruling drew wider notice on November 10, when The Volokh Conspiracy, the legal blog published at Reason, excerpted the opinion.
02
Three accusations in a Great Books course
The opinion recites the facts as the student alleges them, and Bloom reports them that way. She was first accused of using AI in Great Books 191, a classical studies course she took in a fall semester. The graduate student instructor filed an academic misconduct accusation with OSAA, saying he was “immediately suspicious” that her paper was AI-written. She answered that her disability-related writing traits could be misread as AI.
Then she was accused of using AI a second time. That accusation attached what her complaint calls “AI-generated comparison outputs created by prompting AI with [Plaintiff’s] own outline and with instructor feedback.” In plain terms, she says the instructor asked an AI tool to write from her own outline and his own comments, then held the result up against her paper. She calls those “circular AI-comparison outputs generated from Plaintiff’s own materials.”
When she declined a meeting outside the OSAA process, the instructor allegedly described that choice as “self-destructive” and then filed another accusation. She was accused of using AI yet again. Her complaint alleges he had a “paranoid” predisposition to find AI in her work. Those are her allegations. No court has found them true.
OSAA found her responsible on the first two accusations. The sanctions, according to the opinion, were disciplinary probation noted on her unofficial transcript, a required reflective essay and the rewriting of course papers. A grade of “NR” was blocking her degree requirements. For a student accused of using AI, that’s not an abstraction. It’s a transcript and a degree.
03
A complaint, a pause and a third finding
Accused of using AI and sanctioned, the student pushed back through every door she could find. On the same day she appealed, she filed complaints with the U.S. Department of Education and with the university’s Equity, Civil Rights, and Title IX Office, known as ECRT. She also took her concerns to the dean of the College of Literature, Science, and the Arts and to the university president’s office.
OSAA paused her appeal while ECRT reviewed her complaint. ECRT then “dismissed Plaintiff’s complaint given her approved accommodations,” the opinion says. After that, OSAA issued the same misconduct finding on the third accusation. The opinion doesn’t say what became of her complaint to the Department of Education.
The opinion sums up her theory of the case in one line: the defendants “treated disability-related traits and limitations as evidence of guilt and failed to apply disability-informed modifications.” For a disabled student accused of using AI because of how her mind shapes a sentence, that line is the heart of it.
04
Why the disability discrimination claim survived
To weigh the ADA and Rehabilitation Act claims, the court applied the Sixth Circuit’s test from Knox County v. M.Q., a 2023 decision.
An intentional disability discrimination claim requires that the defendant treated the plaintiff less favorably because of her disability. The opinion stresses that “[p]roof of discriminatory motive is critical.” Citing Doe v. BlueCross BlueShield of Tennessee, it also notes that the Rehabilitation Act doesn’t reach disparate-impact claims in the Sixth Circuit.
So for the student accused of using AI, the question was motive. Taking her allegations as true, as a court must on a motion to dismiss, Judge Michelson looked at them together. She pointed to the instructor’s alleged “paranoid” predisposition, his three accusations and the “self-destructive” remark. She also pointed to the student’s claim that OSAA “disregarded the clinical record” she supplied.
Read together, the court found that the student “plausibly pleads that they discriminated against her by disregarding them to find that she used AI.” The word “them” refers to her medical records. That’s the core of the ruling, and it matters for every disabled student accused of using AI.
The claim isn’t simply that she was accused of using AI. Suspicion alone isn’t discrimination. Her claim is that the people judging her had evidence that her disability explains her writing, and set it aside to reach a finding against her. That’s the kind of motive the disability laws ask courts to examine.
The surviving claim has a particular shape. Her ADA and Rehabilitation Act claims against the four individuals in their personal capacities are gone, because she didn’t contest their dismissal. What survives runs against the defendants in their official capacities, together with her request that the court declare her rights.
05
What the court dismissed, and why
The due process claim fell for two separate reasons. First, Judge Michelson held that sovereign immunity shields the university from most of what the student asked for. “Sovereign immunity otherwise protects all other forms of relief,” the opinion says, after recognizing a narrow exception for limited forward-looking relief.
Second, the four individuals get qualified immunity on that claim. The court held that it wasn’t clearly established that an instructor violates due process by “using AI-detection software to analyze that writing and bring an academic misconduct allegation.” On the merits, the court found that she received notice and chances to respond, and that OSAA considered her medical documentation. The court wrote that it “cannot find that Defendants departed from accepted academic norms or failed to exercise professional judgment.”
The failure-to-accommodate claim failed on similar ground. She pointed to two requests: to present her defenses in writing rather than in more live Zoom proceedings, and that her writing style alone not be treated as proof of AI misuse. She conceded the second was accommodated. Because OSAA reviewed her medical records, the court held: “That Plaintiff disagrees with OSAA’s ultimate disposition does not mean OSAA failed to accommodate her by reviewing her medical records.”
That sentence repays a second reading. In the court’s view, disagreeing with the outcome isn’t the same as being denied an accommodation. The accommodation claim and the discrimination claim ask different questions, and only the second survived.
The retaliation claim failed because OSAA paused her appeal under a university policy that applies to all students. “This means the pause in her appeal was standard compliance with policy and not because Plaintiff filed a complaint,” the court wrote.
06
When a careful writing style is mistaken for AI

The court didn’t decide whether AI detection works, and Bloom won’t pretend it did. The opinion records the student’s description of her own writing as her contention, not as a finding. But the fear of being accused of using AI for writing carefully is older than this case, and it reaches far more students than one.
In 2023, Stanford researchers ran seven AI detectors on essays written by people. Stanford HAI reported that the detectors were “near-perfect” on essays by U.S.-born eighth-graders. Yet they “classified more than half of TOEFL essays (61.22%) written by non-native English students as AI-generated.” All seven detectors “unanimously identified 18 of the 91 TOEFL student essays (19%) as AI-generated,” and 89 of the 91 were flagged by at least one.
The article explains why. Detectors “typically score based on a metric known as ‘perplexity,’ which correlates with the sophistication of the writing.” Non-native speakers, it says, tend to score lower on measures such as lexical richness and syntactic complexity. James Zou, the study’s senior author, drew the lesson for schools: “we need to avoid relying on detectors in educational settings, especially where there are high numbers of non-native English speakers.”
That study was about language learners, not disabled writers, and Bloom draws no conclusion from it about this student or this course. What it shows is narrower and still important. A human being’s ordinary way of writing can look like a machine’s to a machine. For a student whose disability shows up as meticulous structure and steady consistency, being accused of using AI can turn a trait of her disability into a piece of evidence against her.
That’s why AI detection belongs in any conversation about disability rights on campus. When a school treats the output of a tool, or an instructor’s hunch, as proof, the students who write differently are the first to be accused of using AI.
07
What this ruling doesn’t decide
The ruling doesn’t decide whether the student accused of using AI actually used it. It doesn’t decide that the university discriminated against her. It doesn’t decide whether AI detection tools are reliable, or when a school may act on an instructor’s suspicion. A motion to dismiss asks only whether a complaint, taken as true, states a claim. The answers to the harder questions come later, if the case reaches them.
It’s also one trial court’s ruling in one case. It doesn’t set a rule for every school, and the university will have its chance to answer the surviving claim.
What it does decide is real. A student accused of using AI can bring a disability discrimination claim when she alleges that decision-makers ignored medical evidence about how her disability shapes her work. For disabled students who’ve watched a careful paper come back with a question mark on it, that matters.
08
Your rights if you’re accused of using AI

Bloom isn’t a law firm and can’t give legal advice. But two federal laws sit at the center of this case, and they reach far beyond Michigan. This student sued under the Americans with Disabilities Act and the Rehabilitation Act. Title II of the ADA is the part that covers state and local governments.
“Under Title II, all state/local governments must follow the ADA regardless of their size,” ADA.gov explains, and it lists “Public education (schools)” among the programs Title II reaches. The same page says governments “can’t deny people with disabilities the chance to participate or make them participate in different programs than available to others.”
If you’re a disabled student accused of using AI, these steps can protect you:
- Save every draft, outline, note and file history you have, with dates.
- Put your disability-related concerns in writing to the office handling the case, and attach your documentation. This student did.
- Ask for any accommodation you need in the process itself, such as answering in writing, and keep a copy of the answer.
- Use your school’s civil-rights office. This student filed with her university’s Equity, Civil Rights, and Title IX Office.
- You can also bring a complaint to the U.S. Department of Education, as she did. For questions about the ADA, the Justice Department’s ADA Information Line is 800-514-0301 (TTY 1-833-610-1264).
- Talk to a disability-rights lawyer or a legal aid office early, before a deadline passes.
None of this guarantees an outcome for anyone accused of using AI. All of it builds a record. In this case, the record was her clinical documentation, and the claim that survived is that the people judging her disregarded it.
09
Oregon students and the same federal laws
Bloom is based in Oregon. A federal trial court ruling in Michigan doesn’t change the law here, and it doesn’t bind Oregon courts. But the ADA and the Rehabilitation Act are federal laws, and the questions in this case will come to Oregon classrooms too.
Students in Oregon who are accused of using AI can ask the same questions this student asked. Did anyone read my documentation? Did anyone ask how my disability shapes my writing? Was the evidence against me anything more than a feeling, or a machine’s guess built from my own outline?
Parents, teachers and disability advocates can ask those questions too, before anyone accused of using AI ends up with a finding on a transcript. You can read more about Bloom’s disability advocacy work.
10
A person, not a probability score
Strip away the case number and the doctrine, and this is a student who says she told her instructors about her accommodations, answered the accusations and handed over her medical records. She says it didn’t matter. A court will decide whether her school listened.
Disabled people know this pattern. We’re asked to prove pain no one can see and needs no one wants to pay for. Now some of us are accused of using AI and asked to prove that our own sentences are ours. A world that’s learning to suspect machines mustn’t learn to suspect disabled people for writing with care.
Meticulous isn’t misconduct. Consistency isn’t a confession. And a student accused of using AI deserves what every person deserves when they’re accused of anything: someone who reads the whole record before deciding.
11
What Bloom is doing
Bloom’s newsroom will follow Doe v. Regents of the University of Michigan as it moves ahead, report what the court decides for the student accused of using AI, and link the filings so you can read them yourself. You’ll find that coverage in the Disability Advocacy newsroom.
On November 9 we reported that the Justice Department sued Lyft under the ADA, another case that tests what federal disability law demands in daily life. We’ll keep reporting both, and we’ll keep naming the courts, the laws and the dates so you can check every word.
And if being accused of using AI has left you frightened, or doubting the way your own mind works, Faith in Bloom offers pastoral care with no condition attached.
12
Frequently Asked Questions
On November 9, 2026, Judge Laurie J. Michelson granted in part and denied in part the defendants’ motion to dismiss. She dismissed the due process, failure-to-accommodate and retaliation claims, but let the intentional disability discrimination claim of the student accused of using AI go forward under the ADA and the Rehabilitation Act.
No. The court decided only that her complaint, taken as true, plausibly states a disability discrimination claim. Whether discrimination happened is still to be decided.
According to the opinion, a graduate student instructor said he was “immediately suspicious” that her paper was AI-written and filed three accusations. One attached outputs made by prompting AI with her own outline and his feedback.
She says she has generalized anxiety disorder and obsessive-compulsive disorder, and that the tone, meticulous structure and stylistic consistency of her writing can be misread as AI.
No. The opinion doesn’t assess the reliability of AI detection. On the due process claim, it held only that no clearly established law barred an instructor from using such software to support an academic misconduct allegation.
Keep every draft and file history, put your disability-related concerns and documentation in writing, and use your school’s civil-rights office. You can also complain to the U.S. Department of Education and talk to a disability-rights lawyer.
This student sued under the Americans with Disabilities Act and the Rehabilitation Act. ADA.gov says all state and local governments must follow the ADA, and it lists public education among the programs Title II covers.
The case continues on the disability discrimination claim of the student accused of using AI against the defendants in their official capacities, along with her request for declaratory relief. The opinion sets no schedule.
Sovereignty over Servitude. ๐ธ
Support this work
Everything Bloom publishes is free to read and free to share.
Bloom Ministries is a registered 501(c)(3) in Salem, Oregon. Monthly giving is what lets us say yes before the money is in hand.


