TL;DR
Get school and study supplies delivered free — and shop member deals
- Fast, free delivery on millions of items
- Access to Prime Big Deal Days deals on October 6–7
- Prime Video, Amazon Music and more included
A guest essay on Astral Codex Ten recounts how a couple used ChatGPT for fertility questions and says the chatbot suggested an MRI that revealed a previously undetected 6-centimeter uterine fibroid. The account describes one couple’s experience, not evidence that AI can reliably diagnose infertility or replace medical care.
A couple struggling with unexplained infertility say ChatGPT suggested an MRI that led doctors to identify a previously undetected 6-centimeter uterine fibroid, according to a guest essay by Drew Housman on Astral Codex Ten. The story illustrates how one patient used an AI chatbot to generate a question for clinicians, but it is a single personal account, not evidence that chatbots can diagnose infertility or safely direct treatment.
Housman writes that he and his wife could create embryos but had experienced repeated unsuccessful embryo transfers. After years of fertility appointments and testing, they had been told their infertility was unexplained. He describes using ChatGPT, including a persona they called “Dr. Reid,” to ask questions about scans, hormone levels, follicle counts, procedures and other parts of treatment that were difficult to cover during limited appointments.
According to the essay, Housman and his wife checked many of the chatbot’s answers with their human doctors and found them consistent. He also describes errors: on rare occasions, he says, the model misread an ultrasound and told them his wife was pregnant. The account does not provide medical records, a systematic review of the chatbot’s accuracy or independent confirmation of how often its advice was correct.
The turning point came when the chatbot suggested that his wife get an MRI. Housman says their fertility doctor was initially reluctant, and the couple had to insist before the scan was ordered. The MRI found a submucosal fibroid about 6 centimeters in diameter. He writes that the clinician reviewing the scan recommended removing it. The essay does not say whether removal took place or report a later pregnancy outcome.
A Prompt for Further Medical Testing
The account matters because it shows one way a patient-facing chatbot can affect a medical conversation: it raised a test the couple says had not been pursued during years of care. The MRI then identified a finding that their fertility doctor considered important enough to address. In this case, the AI did not order the scan or perform the diagnosis; the couple brought its suggestion to a clinician, who ultimately authorized the imaging.
It also highlights the limits of treating a striking personal story as proof. The essay cannot show that ChatGPT is generally able to spot missed causes of infertility, that MRI would be suitable for other patients, or that the fibroid caused the failed embryo transfers. Those questions require clinical evidence and an individual medical assessment. For readers, the narrower takeaway is that AI-generated questions may help some patients prepare for appointments, while decisions about tests and treatment remain matters for qualified clinicians.
As an affiliate, we earn on qualifying purchases.
Six Years of Unexplained Infertility
Housman describes a lengthy fertility journey involving multiple doctors, scans, tests and biopsies. He and his wife could create embryos, but transfers did not result in a pregnancy. In the essay, he characterizes “unexplained infertility” as a label used after clinicians have not identified a cause; he says that after receiving it, the couple felt that doctors were less actively investigating their case.
The pair began consulting ChatGPT after GPT-4 became available, using it to pursue detailed questions outside their appointments. Housman says early responses were sometimes blocked by the system’s medical-safety restrictions. He recounts prompting it to answer as a fictional television doctor and later calling the resulting persona Dr. Reid. These descriptions are the author’s recollection of using the chatbot; they do not establish how current AI systems perform or how they should be used in medical care.
The essay’s central episode is the contrast between the couple’s extensive prior testing and the MRI recommendation. Housman says a medical note had mentioned MRI once as a possibility, but the option was not emphasized and did not lead to a scan. The later imaging found the fibroid. The source does not include a detailed clinical timeline, the full test history or the doctors’ explanations for their earlier decisions.
“That’s when our LLM doctor stepped in.”
— Drew Housman, guest essay on Astral Codex Ten
The Outcome and Medical Evidence
The essay does not report whether the fibroid was removed, whether the couple later became pregnant, or whether doctors concluded that the growth explained the failed transfers. It also does not include the clinicians’ account of why MRI was not pursued earlier, or the full imaging and treatment records needed to assess the sequence independently.
More broadly, the story offers no controlled comparison of chatbot advice with specialist care and no estimate of how often similar systems make errors. Housman himself recounts at least one kind of significant misreading, when the chatbot interpreted an ultrasound as showing pregnancy. The source therefore supports reporting what happened to this couple as they describe it, but not a general claim that AI can identify missed diagnoses or improve fertility outcomes.
Clinical Follow-Up Remains Unreported
The next development in the couple’s medical story is not provided in the essay: it does not say whether surgery occurred or what happened after the fibroid was found. No further clinical milestone, follow-up statement from their fertility team or later outcome is included in the supplied source.
For readers considering AI for health questions, the account supports a limited approach: use chatbot responses as prompts to discuss with a clinician, not as a diagnosis or substitute for examination and medical judgment. The source does not describe a formal study or policy change arising from the experience. Any broader conclusions about AI-assisted fertility care would require clinical research beyond this single report.
Key Questions
What did the MRI find?
According to Drew Housman’s account, it showed a submucosal uterine fibroid about 6 centimeters in diameter that had not previously been detected.
Did ChatGPT diagnose the fibroid?
No. Housman says ChatGPT suggested that his wife get an MRI. The scan identified the fibroid, and a fertility doctor reviewed the result and recommended removal, according to the essay.
Does the account prove AI can diagnose infertility?
No. It is a personal account of one couple’s experience, not a clinical trial or accuracy study. It cannot establish how reliably AI can identify medical conditions for other patients.
Did the couple have the fibroid removed or later become pregnant?
The supplied essay does not report whether the fibroid was removed or provide a later pregnancy outcome.
Source: rss
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
