A boy saw 17 doctors over 3 years for chronic pain. ChatGPT found the
diagnosis
...
...
In total, they visited 17 different doctors over three years. But Alex
still had no diagnosis that explained all his symptoms. An exhausted
and frustrated Courtney signed up for ChatGPT and began entering his
medical information, hoping to find a diagnosis.
...
...
She eventually found tethered cord syndrome and joined a Facebook
group for families of children with it. Their stories sounded like
Alex's. She scheduled an appointment with a new neurosurgeon and told
her she suspected Alex had tethered cord syndrome. The doctor looked
at his MRI images and knew exactly what was wrong with Alex.
...
...
There are both free and paid versions of ChatGPT, and the latter works
much better than the free version, Beam says. But both seem to work
better than the average symptom checker or Google as a diagnostic
tool. "It's a super high-powered medical search engine," Beam says.

It can be especially beneficial for patients with complicated
conditions who are struggling to get a diagnosis, Beam says.

https://www.today.com/health/mom-chatgpt-diagnosis-pain-rcna101843

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

There are both free and paid versions of ChatGPT, and the latter works
much better than the free version, Beam says. But both seem to work
better than the average symptom checker or Google as a diagnostic
tool. "It's a super high-powered medical search engine," Beam says.

It can be especially beneficial for patients with complicated
conditions who are struggling to get a diagnosis, Beam says.

https://www.today.com/health/mom-chatgpt-diagnosis-pain-rcna101843

Cool! This however brings up an interesting question: ideally you'd
want to nourish ChatGPT with billions of medical records, diagnoses,
and transcripts. Since all that is protected, confidential, and
not-shareable information, what did ChatGPT get trained on?

I want access to a ChatGPT that mowed through every known clinical
record and doctor's report, not a ChatGPT that got trained on the kind
of garbage you typically find on the open sewer of the Internet. I
learned long ago, when looking up health info, to rely only on
Mayoclinic and one or two others for actual, factual information. The
rest of it is unfiltered trash, much of it written explicitly to
capture searches like "do avocados raise your cholesterol level" and
sell a few ad impressions. Toxic and stupid at best, misleading and
dangerous at its worst.

Postscript: Tethered cord syndrome looks terrible: https://www.aans.org/en/Patients/Neurosurgical-Conditions-and-Treatments/Tethered-Spinal-Cord-Syndrome

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Tue, 12 Sep 2023 19:43:21 -0400, Retrograde <[email protected]>
wrote:

I want access to a ChatGPT that mowed through every known clinical
record and doctor's report, not a ChatGPT that got trained on the kind
of garbage you typically find on the open sewer of the Internet.

IBM Watson Health is now Merative https://www.ibm.com/watson-health/merative-divestiture

==============================

But it's not likely to replace a clinician's expertise anytime soon,
he says. For example, ChatGPT fabricates information sometimes when it
can't find the answer. Say you ask it for studies about influenza. The
tool might respond with several titles that sound real, and the
authors it lists may have even written about flu before -- but the
papers may not actually exist.

This phenomenon is called "hallucination," and "that gets really
problematic when we start talking about medical applications because
you don't want it to just make things up," Beam says.

https://www.today.com/health/mom-chatgpt-diagnosis-pain-rcna101843

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

JAB <[email protected]d> writes:

A boy saw 17 doctors over 3 years for chronic pain. ChatGPT found the diagnosis
...
...
In total, they visited 17 different doctors over three years. But Alex
still had no diagnosis that explained all his symptoms. An exhausted
and frustrated Courtney signed up for ChatGPT and began entering his
medical information, hoping to find a diagnosis.
...

[snip]

The connectionist revival took of in the mid 80s. Already in the early
90s, hand-hacked neural nets trained on specific areas of medicine
(and incredibly feeble by today's standards) were already achieving
diagnostic success equal to or in some cases better than human
specialists. Pattern recognition is the underlying magic of neural
nets and clincal diagnosis is a perfect match given good training
data.

My take is that is that the people developing the currently hyped
chat-bots are striving for language excellence -- all the things they
try to teach you in high school English, uni-level composition,
journalism school and the like. Semantics, factual accuracy,
substantive insight on a topic and other desirable qualities aren't
sexy and are getting left as an exercise for somebody else.

https://www.today.com/health/mom-chatgpt-diagnosis-pain-rcna101843

--
Mike Spencer Nova Scotia, Canada

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 21 Sep 2023 02:31:45 -0300, Mike Spencer
<[email protected]e> wrote:

neural nets trained on specific areas of medicine
....achieving diagnostic success

Med school presents the top 40 conditions, so to speak, since
thousands more exist.

Something like Radial Neuropathy (Saturday Night Palsy) could be
mistaken for a stroke if not covered in school, for instance.

https://precisionemg.com/radial-neuropathy-saturday-night-palsy/

chat-bots are striving for language excellence

Gmail, when composing an email, attempts to correct grammer
============

Using chatbots for English language learning in higher education
...
...
Despite the students' disagreements with the social influence of
Chatbots on their behavioural intentions, Chatbots are still
considered beneficial enough to serve as an interlocutor for English
language learning.

https://www.sciencedirect.com/science/article/pii/S2666920X23000322.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)