Re: Be very, very afraid...

On 10/28/2024 3:10 PM, Lasse Langwadt wrote:

On 10/27/24 02:43, Don Y wrote:

No, not that AI is going to take your job ("render you
redundant") but, rather, that it is going to be relied upon
for information that is provably incorrect:
>
<https://apnews.com/article/ai-artificial-intelligence-health-business-90020cdf5fa16c79ca2e5b6c4c9bbb14>
>
it seems the most expeditious way to get these sorts of problems
fixed is NOT to regulate how it can be used but, rather, to
hold firms using it financially (criminally?) accountable
for its follies...

 https://www.washingtonpost.com/travel/2024/02/18/air-canada-airline-chatbot-ruling/

Imagine an AI interpreting a CT scan and coming to a conclusion
on which a human doctor would otherwise have "hedged".
If the doctor *overrules* the AI and the pt suffers some "loss",
he's screwed.
If he *defers* to the AI and the pt suffers a loss, same outcome.
As to the originally cited issue, who's responsibility is it to
ensure the accuracy of the transcription of a pt visit?  Prior
to AI, the pt would just see a *summary* of his visit -- no specific
notes from the provider.
Will the pt have to vouch for the transcription?  Will the provider?
It's a standing joke, here, that providers always want you to
show up 15 minutes early -- "for paperwork":
     "Review these documents and make any corrections that are necessary..."
Yet, the corrections are still not incorporated for your NEXT visit.
Ah, here's a provider that lets you review the documents
ELECTRONICALLY and make corrections to them "live"!  Shirley,
these updates will find their way into the record...?
Nope.
"Look, I have told you EACH TIME I VISITED about my allergies
and contraindicated medications.  Yet, each visit shows that
section as 'blank'... is there a problem with YOU system?
Perhaps I should ask YOU to initial my corrections so I have
more conclusive proof if you screw the pooch?"