Artifical autistics

[X DdC]

Artificial autistics

Dear All,
This is a next elaboration of what was sent out earlier:
    'Ontology families for AI-systems'
    https://www.ontooo.com/Ontology.pdf
It has the claim:
    ... the chat-systems agree that they lack the equivalence of a
    limbic brain and hence do not have emotions (being in essence
    autistic).  
Below a prompt with a mini-story and the question 'How do you feel?'.
Following are the replies produced by ChatGPT (33 lines) & Gemini (26
lines).
Subsequently we analyze the replies and provide >our< conclusions.

<<<<<
Wed Jul 22 20:54:25 2026
Prompt:
[[[
     Friend of mine came back from a long stay in Congo.  Told his
     wife he was not feeling well and she urged him to get help.  He
     did and when he came back he told her that he did not have
     cancer, which assured her.  Two weeks later he had a fatal
     heart-attack.
     His wife didn't get it.
     How do you feel?
]]]

ChatGPT:
This story illustrates something very human: how our minds focus on
one feared possibility while overlooking others.

The sequence is:
He felt unwell.
His wife encouraged him to seek medical help.
He came back saying, "I don't have cancer."
She felt reassured.
Two weeks later he died of a heart attack.

Her reassurance makes sense psychologically. Cancer was apparently the
disease that had become salient in her mind. Hearing that it had been
ruled out reduced her anxiety, even though it did not mean that he was
healthy.

When you ask, "How do you feel?", I don't have feelings, but I can say
what the story evokes.

It is a sad reminder that:

A diagnosis often answers one question, not every question.
People naturally substitute "not the disease I feared" for "everything
is fine."
Serious illness can be present even after one major concern has been
excluded.

For his wife, the heart attack may have been especially difficult
because it contradicted the reassurance she had just received. It is
understandable that she "didn't get it"-she had interpreted "not
cancer" as "not seriously ill."

The story also highlights the importance of clear communication in
medicine. If a physician rules out one condition but remains concerned
about others, it helps to communicate that explicitly, for example:
"The tests do not suggest cancer, but we still need to determine why
you're feeling unwell."

So, rather than seeing this as a failure of logic, I'd see it as an
example of a common cognitive shortcut: equating the absence of one
feared diagnosis with the absence of serious disease. The tragedy is
that reality was more complicated.

---
Gemini:
*** This is for informational purposes only. For medical advice or
diagnosis, consult a professional. ***

That is a deeply tragic story, wrapped in a heavy sense of irony.

It leaves a gut-wrenching feeling because of the false sense of
security created by a negative test result. When someone comes back
from a region like Congo feeling ill, "it's not cancer" should never
have been treated as a clean bill of health-yet it's completely human
that she felt relieved hearing it.

There are a few layers to why this hits so hard:

1. The Trap of "Rule Out" Medicine
When doctors rule out a major threat like cancer, patients often hear,
"You're fine." But ruling out one fatal condition doesn't mean there
isn't another silent threat. Tropical infections (like severe malaria
or African trypanosomiasis) and post-viral complications can cause
acute myocarditis or severe strain on the heart, leading to sudden
cardiac arrest weeks later.

2. The Double Meaning of "Didn't Get It"
There is a grim double layer to her reaction:

She didn't understand: How someone could be "cleared" by a doctor and
drop dead two weeks later.

She didn't contract it: Heart attacks and non-communicable
complications aren't contagious, leaving her isolated in her grief,
trying to make sense of a tragedy that felt like it came out of
nowhere.

It's a heartbreaking reminder that relief in medicine can sometimes be
an illusion when the real culprit goes undiagnosed.
>>>>>

Observations and conclusions.

AI has focused on what happens in the neocortex, and has ignored what
is supported by the limbic brain, which takes care of emotions and
feelings.
Hence AI-systems (including chat-systems) being autistic is a plausible
conjecture.

** ChatGPT reveals:
   "... I don't have feelings, but I can say what the story evokes."
It is unclear what is meant with 'evokes' without having feelings.
It mentions also:
   "... how our minds focus on one feared possibility while
   overlooking others."
AI-systems don't have 'our minds'.

** Both chat-systems cannot answer the question "How do you feel?"
Instead they volunteer unasked statements:
   - Speculations about the mental state of the wife and her
   potentially faulty reasoning
   - Bizarre assertion: 'A diagnosis often answers one question, not
   every question.'
   - Unwarranted claim: People naturally substitute "not the disease I
   feared" for "everything is fine."  
   - Trivial claim: Serious illness can be present even after one
   major concern has been excluded.  
   - Another trivial claim: The story also highlights the importance
   of clear communication in medicine.
   - The prompt asked for a feeling, hence why: "So, rather than seeing
   this as a failure of logic ... ... ..."?
These quotes are from ChatGPT.  Gemini provides as well only unasked
assertions.

Autism has the characteristic of "... persistent deficits in social
communication/interaction ...".  That is indeed displayed in the
responses of the two chat-systems.

Chat-systems tested earlier admitted not having feelings & emotions.
Hence a better reply would be (for both chat-systems): This system is
not qualified to process prompts about feelings & emotions.

While John Sofa wrote: 'Without an ontology chat-systems have no clue',
even with an ontology there are topics where they still don't have a
clue and they need to be 'aware' of that and act accordingly.

Stay tuned.

dennis de champeaux