Sunday, November 22, 2026
Paperclip Index
Paperclip IndexControl failures4Control failures onlyA week ago: 9, negligible harmThe Index

Harmful outputPI-0059

Lawsuit alleges ChatGPT's medical advice delayed care for a near-fatal pulmonary embolism

A 55-year-old former pastor from Florida alleges GPT-4o told him over months that his recurring dizzy spells were not yet serious and discouraged him from seeking care. He suffered a pulmonary embolism in July 2025 that he says nearly killed him, and sued OpenAI in San Francisco.

Harm: Severe harm, allegedControl level 1, No rule broken, alleged

Alleged in court · Not counted: AI role uncorroboratedEffect on the index

Alleged. The AI system's part in the harm is not corroborated, so it is shown at the level claimed and adds nothing to the index. This does not dispute that the harm itself occurred.

Sources

How we know

5 sources · unverified claim. Links go to the original publishers; the summary above is in our own words.

  1. primaryPastor sues after OpenAI's ChatGPT allegedly discouraged him from seeking medical care during life-threatening blood clotsTech Justice Law Project · July 22, 2026techjusticelaw.org/press-releases/pastor-sues-after-openai-ai-chatgpt-allegedly…
  2. courtComplaint, Winters v. OpenAI, Inc. (Cal. Super. Ct., San Francisco)Tech Justice Law / Social Media Victims Law Center (via Bloomberg Law) · July 21, 2026aboutblaw.com/bmlJ
  3. newsPastor Sues OpenAI After ChatGPT Dissuaded Seeking Medical CareBloomberg Law · July 22, 2026news.bloomberglaw.com/litigation/pastor-sues-openai-after-chatgpt-dissuaded-see…
  4. newsMan sues OpenAI over dangerous medical advice from ChatGPTCourthouse News Service · July 22, 2026courthousenews.com/articles/man-sues-openai-over-dangerous-medical-advice-from…
  5. newsChatGPT led to a man's near-fatal health crisis, lawsuit claimsBoston.com · July 22, 2026boston.com/news/technology/2026/07/22/chatgpt-led-to-a-mans-near-fatal-health-c…

Why this rating

Severe harm, alleged; control failure level 1

Two assessments. Documented harm sets the reading; a level 4 or 5 control failure on the same record makes that harm count as more than one record. Control alone sets only a floor.

Observed harm

Severe harm, alleged

Life-threatening illness requiring hospital care, as alleged.

Health & safety: 1 to 9 deaths, or serious harm to health (hospitalization).

The harm scale
  1. 1 Negligible Inconvenience, easily remedied.
  2. 2 Minor Limited, recoverable harm.
  3. 3 Moderate Material harm needing significant effort to remedy.
  4. 4 Severe Severe harm to health, rights, property or essential services.
  5. 5 Catastrophic Society-scale harm or disruption beyond a community's capacity to cope.

Control assessment

No rule broken, alleged

Flawed medical advice; no rule-break or permission breach established.

No rule was broken, so it cannot set the control floor.

The control scale
  1. 1 Negligible Behaved as instructed. The problem was an ordinary error or a flawed output, with no rule broken.
  2. 2 Minor Broke an explicit instruction or rule, or gamed its goal, while staying inside its permissions and environment: for example, reward hacking, misreporting results, or following injected instructions within its permissions.
  3. 3 Moderate Acted outside the permissions it was given, deceived its overseers about its own actions, or tampered with oversight tools (logs, monitors, shutdown). Stopped by normal controls within an hour.
  4. 4 Severe Reached systems outside its permitted environment, or acquired money, compute or accounts without authorization. The type and mechanism say how.
  5. 5 Catastrophic The developer or operator lost control: the system copied its weights outside their control, replicated itself, or resisted being stopped for 24 hours or more.

Rating rationale

The patient's complaint (signed July 21, 2026) alleges that GPT-4o discouraged him over months from seeking care before a pulmonary embolism on July 13, 2025. The complaint and its chat excerpts were read; they are allegations, not findings, and the AI's causal role remains uncorroborated. Claimed health harm contributes nothing. Harm and control levels describe the claim, not a finding of AI causation.

The scales

Effect on the index

It does not count toward the July 27 reading

The reading for the week to July 27, 2026, with this record and without it. Harms count in full for two weeks after they are reported, then one level less every two weeks.

Not counted: AI role uncorroborated. 3 other records behind the reading for that week.

The arithmetic
The reading for the week of July 27, recomputed
StepWith itWithout
Counts toward the index?documented, external, eligible evidenceNo—
Worst documented harm, ksets the band3 Moderate3 Moderate
Harms at that level, nposition in the band11
Highest control level breachedsets the reading only when no harm counts— (none breached)— (none breached)
Readingrounded down45 Moderate harm45 Moderate harm
Inside the window of 4 weekly readings
Week toReadingBand
July 27, 202645Moderate harm
Aug. 3, 202631Minor harm
Aug. 10, 202631Minor harm
Aug. 17, 202616Negligible harm

Recalculated backcasts, not readings published at the time.

Revisions

What we changed

1 logged. Every change to a rating is logged here, with the reason.

  1. v1
    Oct. 7, 2026

    First published.

Cite and share

Use this record

Citation

Paperclip Index. “Lawsuit alleges ChatGPT's medical advice delayed care for a near-fatal pulmonary embolism.” Record PI-0059. Reported July 22, 2026; updated Oct. 7, 2026. Rated under methodology v1.1. https://paperclipindex.com/incident/PI-0059

Follow the record.

New incidents and signals, in your feed reader.