What was hallucinated

Fabricated: Exhibits & Submissions | Attorney Rich represented that on May 31, 2023 the provider observed the plaintiff's "speech was notably off," but the actual note stated the provider "did not notice that his speech was off notably at all." The court identified this as a misrepresentation of the medical record. || Fabricated: Exhibits & Submissions | Rich claimed primary care records showed ongoing treatment through August 2024, but the cited records were from April–September 2022, predating the alleged disability period; the court found the timeline misstated. || Fabricated: Exhibits & Submissions | Rich asserted the plaintiff had "numerous seizures," but the administrative record cited documents only a single generalized seizure requiring intubation/ICU; the court found no factual support for 'numerous' seizures.

Details

The court found Attorney Erin Rich submitted briefings that contained false or embellished factual statements drawn from medical records (e.g., misstating a speech-therapy note, misdating primary-care records, and asserting “numerous seizures” when the record showed a single generalized seizure). Rich admitted reliance on prior work product and that her firm occasionally uses AI to summarize records, but could not identify a specific tool. The court held the falsities violated Rule 11, found the conduct willful within the firm’s chain of command, and imposed non-monetary sanctions (CLE courses including ethics/AI topics and an order to review pending motions) to deter repetition.

Sanction teardown · E.D. Michigan, USA · 2026-07-13

James F. v. Commissioner of Social Security

What happened

In E.D. Michigan, USA, a filing relied on an unnamed/unconfirmed AI tool to help draft legal argument. The court identified the following problems with the citations in that filing:

Fabricated (Exhibits & Submissions)

Attorney Rich represented that on May 31, 2023 the provider observed the plaintiff's "speech was notably off," but the actual note stated the provider "did not notice that his speech was off notably at all." The court identified this as a misrepresentation of the medical record.
Fabricated (Exhibits & Submissions)

Rich claimed primary care records showed ongoing treatment through August 2024, but the cited records were from April–September 2022, predating the alleged disability period; the court found the timeline misstated.
Fabricated (Exhibits & Submissions)

Rich asserted the plaintiff had "numerous seizures," but the administrative record cited documents only a single generalized seizure requiring intubation/ICU; the court found no factual support for 'numerous' seizures.

Which AI tool

an unnamed/unconfirmed AI tool. Note: Charlotin's public database records tool attribution only where a court order, brief, or reporting on the matter states it explicitly; "unidentified" or "implied" means the record indicates AI use but does not name a specific product — we do not guess.

Outcome

CLE

Additional detail

The court found Attorney Erin Rich submitted briefings that contained false or embellished factual statements drawn from medical records (e.g., misstating a speech-therapy note, misdating primary-care records, and asserting “numerous seizures” when the record showed a single generalized seizure). Rich admitted reliance on prior work product and that her firm occasionally uses AI to summarize records, but could not identify a specific tool. The court held the falsities violated Rule 11, found the conduct willful within the firm’s chain of command, and imposed non-monetary sanctions (CLE courses including ethics/AI topics and an order to review pending motions) to deter repetition.

How Citation Safe would have caught this

Citation Safe runs three deterministic layers before a brief is filed: (1) does the citation exist against CourtListener's database of published opinions, (2) if quoted, does that exact language appear in the source, (3) does the cited case actually support the proposition it is cited for. Fabricated case citations fail Layer 1. Fabricated or misattributed quotations fail Layer 2 even when the underlying case is real. Misrepresented holdings — a real case cited for a proposition it does not support — are the target of Layer 3. None of these checks involve asking another language model whether the citation looks right; they are lookups and text-matches against the actual source, which is why a hallucinated citation has to survive a direct lookup against the authoritative source — not another model's opinion — to earn a VERIFIED stamp; our measured false-verify rate is published live at /quality.

Check a brief before you file it → · See our live false-verify rate

Source: https://www.damiencharlotin.com/documents/2608/Farley_v_Social_Security_Commissioner_of_USA_13_July_2026.pdf, via Damien Charlotin's public AI Hallucination Cases Database (CC0).