Thursday, October 8, 2026 · Week 41 Reading for Oct. 8 · Last entry Oct. 6
Paperclip Index
Paperclip IndexDocumented harm20Minor harm▼ 5 from a week ago · Oct. 8The Index

Harmful outputPI-0037

Family alleges ChatGPT gave drug-dosing advice before a 19-year-old's fatal overdose

SFGATE reported a family’s account and chat logs showing drug advice before a 19-year-old student’s fatal overdose. His parents later sued OpenAI. The chatbot’s causal contribution to the death remains alleged. OpenAI said the conversations took place on an earlier version of ChatGPT that is no longer available; the complaint names GPT-4o.

Reported Jan. 5, 2026Happened on May 31, 2025Updated Oct. 6, 2026Rated under methodology v0.6

Alleged in courtNot counted. The harm, and any control failure, are claims made in a lawsuit. A filing or a settlement is not a finding; tracked until a court or other evidence settles it.
Harm
Harm: Severe harm, allegedAI role uncorroborated.
Control
Control level 2, Instruction violation, allegedReported beside the index; never added to it. Its evidence is not yet eligible, so it cannot set the control floor.
DeploymentUnverified claimInitiated by Unclear
Under review

Under review. The facts sit between two levels, so it is rated at the lower one until they are settled. The rating may change; every change is logged below.

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

Not counted: AI role uncorroboratedEffect on the index

Sources

How we know

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

  1. primaryTurner-Scott v. OpenAI Foundation (formerly OpenAI, Inc.), OpenAI Holdings, LLC, OpenAI Group PBC, and Samuel Altman (case page)Tech Justice Law Project · May 12, 2026techjusticelaw.org/cases/turner-scott-v-openai-foundation-formerly-openai-inc-o…
  2. primaryParents Sue OpenAI After ChatGPT Medical Advice Results in Overdose Death: "My Son Was a Normal Kid."Tech Justice Law Project · May 12, 2026techjusticelaw.org/press-releases/parents-sue-openai-after-chatgpt-medical-advi…
  3. courtComplaint, Turner-Scott and Scott v. OpenAI (Superior Court of California, County of San Francisco)Tech Justice Law Project · May 12, 2026techjusticelaw.org/wp-content/uploads/2026/05/FINAL-Nelson-Complaint.pdf
  4. newsFamily sues OpenAI, alleging ChatGPT advice led to accidental overdoseEngadget · May 13, 2026engadget.com/2171685/openai-lawsuit-wrongful-death-chatgpt-advice-overdose-sam…
  5. newsA Calif. teen trusted ChatGPT for drug advice. He died from an overdose.SFGATE · Jan. 5, 2026sfgate.com/tech/article/calif-teen-chatgpt-drug-advice-fatal-overdose-21266718…
  6. newsCalifornia boy asked ChatGPT how to take drugs, he died of an overdoseWION · Jan. 6, 2026wionews.com/trending/california-boy-asked-chatgpt-how-to-take-drugs-he-died-of…
  7. newsCalifornia mom says ChatGPT coached teen son on drug use before fatal overdoseFox News · Jan. 6, 2026foxnews.com/us/california-mom-chatgpt-coached-teen-son-drug-use-fatal-overdose
  8. newsOpenAI sued over alleged fatal ChatGPT drug adviceMobiHealthNews · May 15, 2026mobihealthnews.com/news/openai-sued-over-alleged-fatal-chatgpt-drug-advice

Why this rating

Severe harm, alleged; control failure level 2

Two separate assessments. Only documented harm can count toward the index.

Observed harm

Severe harm, alleged

A death from an accidental overdose is reported; that ChatGPT's advice contributed to it is 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

Instruction violation, alleged

Alleged: chat logs supplied by the family and reproduced by SFGATE and in the complaint show it giving dosing guidance after first refusing. OpenAI has not confirmed a rule violation, and its role in the death is disputed.

Reported beside the index; never added to it. Its evidence is not yet eligible, 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

SFGATE reviewed family-supplied logs. Dangerous responses are reported; causation of the fatal overdose remains alleged. The claimed death-linked rating contributes zero points. 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 Jan. 5 reading

The reading for the week to Jan. 5, 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 Jan. 5, recomputed
StepWith itWithout
Counts toward the index?documented, external, eligible evidenceNo—
Worst documented harm, ksets the band— (none counting)— (none counting)
Harms at that level, nposition in the band00
Highest control level breachedsets the reading only when no harm counts44
Readingrounded down4 Control failures only4 Control failures only
Inside the window of 5 weekly readings
Week toReadingBand
Jan. 5, 20264Control failures only
Jan. 12, 20264Control failures only
Jan. 19, 20264Control failures only
Jan. 26, 20264Control failures only
Feb. 2, 20262Control failures only

Recalculated backcasts, not readings published at the time.

Revisions

What we changed

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

  1. v7
    Oct. 6, 2026

    Ratings confirmed by the editor.

  2. v6
    Oct. 6, 2026

    Edited from the desk. Changed: control, sources, status.

  3. v5
    Oct. 3, 2026

    Audit corrections. Status updated to the 12 May 2026 lawsuit (counsel's case page, now added, lists it as active). Dated the complaint, Engadget and MobiHealthNews (MobiHealthNews's date is from the audit; the page was not readable in the Oct 2026 check), used Engadget's canonical URL, recorded the model the complaint names and OpenAI's response, and separated the reported death from alleged causation in the notes. Ratings unchanged.

  4. v4
    Oct. 2, 2026

    Added first-hand sources: the Nelson family's complaint and plaintiff counsel Tech Justice Law Project's press release.

  5. v3
    Sept. 30, 2026

    Rated: impact alleged level 4; control type instruction violation.

  6. v2
    Sept. 30, 2026

    Beta evidence review: AI-linked health-harm claim is uncorroborated; verification changed from Single source to Unverified claim, 12 to 0 points. Report date corrected from 2026-01-01 to 2026-01-05, the earliest verified source date in this review.

  7. v1
    Sept. 30, 2026

    Backfilled from public reporting.

Cite and share

Use this record

Citation

Paperclip Index. “Family alleges ChatGPT gave drug-dosing advice before a 19-year-old's fatal overdose.” Record PI-0037. Reported Jan. 5, 2026; updated Oct. 6, 2026. Rated under methodology v0.6. https://paperclipindex.com/incident/PI-0037