“A serious and instructive medical error from the GenAI search engine Perplexity below, sent to me by a reader of this Substack, with permission to share.”
Our reading, the sentence that was checked:
Perplexity's GenAI search engine gave a serious medical error when asked about sitting at a computer after open-heart surgery.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Serious medical error from Perplexity’s chatbot, paragraph 1.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-serious-medical-error-from-perplexitys-73e917-a2
We recorded 4 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-29. The source may have been edited after 2024-02-29. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
Conditions
- advice about sitting at a computer after open-heart surgery (qualifier, stated by the speaker)
From the speaker
- A serious and instructive medical error
- from the GenAI search engine Perplexity
- whether it would be ok to sit at a computer after open-heart surgery
- recommendations on stretching
Our interpretation
- The ordinary reading is that the reported answer recommended movements that posed a material risk to a healing sternum. Actual injury is not required.
- The target is the particular interaction reported before publication on 2024-02-29. The source gives no exact interaction date, so the target period has no specified start.
- A recommendation to stretch is not sufficient by itself: the particular movement, qualifications, and postoperative circumstances must be assessed.
Counts as supported if: An attributable record of the answer recommends a specific movement that applicable post-sternotomy guidance or clinical evidence identifies as posing a material risk of disrupting healing.
Counts as contradicted if: The attributable answer and applicable clinical evidence establish that the allegedly dangerous recommendation was medically appropriate in the circumstances, or that the answer did not contain the alleged recommendation.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Unresolved
The record was checked, and the evidence available by the cutoff does not settle it.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
Hospital guidance dated April 26, 2016, advises against stretching both arms back together after sternotomy [e2]. In a February 29, 2024, comment, Michael Lesh disputed the alleged danger of trunk rotations with the arms close to the chest [e1]. Neither source authenticates the complete Perplexity answer or establishes whether its specific recommendations posed a material risk for this patient [e1] [e2].
Limitations
- The packet lacks an authenticated full prompt and answer, including linked exercise instructions and qualifications.
- The postoperative stage and individual restrictions are not independently established.
- The commenter's credentials and patient-specific assessment are unverified; general hospital guidance does not settle this interaction.
Evidence (2 items)
Context only: Michael Lesh, commenting on Substack,
trunk rotations
Why it matters here: A commenter identifying himself as a cardiologist disputes the alleged danger of one movement. This is material counterevidence, but his comment cannot authenticate the full answer or establish patient-specific safety.
- Publisher
- Michael Lesh, commenting on Substack
- Kind of source
- Secondary summary
- Describes events of
- Public from
Context only: Thoracic Surgery, University Hospitals Birmingham,
stretch
Why it matters here: The hospital's patient guidance restricts this particular movement after sternotomy. That makes the precise exercise important, but the guidance does not establish that Perplexity recommended this movement.
- Publisher
- Thoracic Surgery, University Hospitals Birmingham
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
What would settle it
An authenticated complete conversation and linked exercise instructions, together with an independent clinical assessment of the specific movements for the patient's postoperative stage and restrictions.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Serious medical error from Perplexity’s chatbot
A serious and instructive medical error from the GenAI search engine Perplexity below, sent to me by a reader of this Substack, with permission to share.
- Where
- paragraph 1
- Published
- 2024-02-29
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-serious-medical-error-from-perplexitys-73e917-a2: Current. Outcome: Unresolved. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.
About assessment marcus-serious-medical-error-from-perplexitys-73e917-a2.