As far as I’m concerned, using AI of any sort inside of a doctor’s office is a grave violation of HIPAA
Australia’s peak regulator of medical practitioners, AHPRA, said clinicians must always check all output from an AI scribe for accuracy to ensure they meet their professional obligations.
Show me one single AI user that checks the output of their AI, and I will show you my unicorn
I read “medieval” appointment and was thoroughly confused, and intrigued.
deleted by creator
I’ve used AI for medical transcription before and it is absolutely incredible. It’s not just a word for word text, but a detailed clinical breakdown of everything discussed.
All you have to do is simply read the damn thing thoroughly and correct whatever minor errors pop up.
Takes 95% less time.
All you have to do is fire the human who is trained and rarely makes a mistake and replace them with a paid machine that almost always makes a mistake and double check everything yourself (or hire another human to double check it) because someone’s life literally depends on the entirety of the text being absolutely correct.
AI does work at a level that makes unpaid interns look like career professionals. All for the low cost of completely fucking the environment.
Glad you like it.
I’m unsure if there’s an ethical issue with putting patients medical records in the data machine.
Medical AI has guardrails out the wazoo. By the time it’s made available to a physician, it’s probably been reviewed by a council, vetted by IT, and negotiated heavily by Legal. If the data leaves the premises, it’ll be thoroughly protected and isolated.
That said, fuck AI. I’d give up medical transcription to get the environment, economy, and sanity back. Musk’s data center is running dozens of natural gas turbines in a low-income, primarily Black neighborhood 11 miles down the road from St. Jude Children’s Research Hospital. It’s positively ghoulish.
It needs to be a system and a contractual framework that complies with the medical privacy regulations of the place where it’s being used (in the US, HIPAA). I have some doubts about whether specific companies or people will actually comply, or the enforcement of those laws by certain governmental authorities, but I don’t think it’s conceptually impossible to have this kind of tech be compliant with those types of regulations.
There exists online and offline ML.
Transcription is not a task that needs a full blown LLM, summarization probably does, and if being done offline or on a server with trusted verifiable lack of logging should be fine. That does not mean send it to ChatGPT API.
I would hope they store the raw conversation file encrypted with patient records so they can retranscribe later as tools improve.
Exactly. It is a tool for professionals to use to speed up existing processes. It is not a replacement for people, regardless of what the marketers tell you so you buy their product. I don’t understand how there aren’t thousands of lawsuits against these AI companies for their blatantly false advertising.
AI keeps proving this day by day when people try to make it a replacement and have it blow up in their faces every day.
blatantly false advertising.
SEC is neutered, what makes you think anyone in charge of advertising rules hasn’t been replaced by a 17 year old called Goebbels Gobbler or something with a DOGE laptop.
SEC is neutered
This article is about an incident in Australia. The SEC has no bearing on what Australia can do.
Hell, I just found out a few months ago my father’s office (he’s a lawyer) still did transcriptions of court proceedings manually! Full blown lawyers wasting expensive man hours doing that. As soon as I found out I got some random whisper cpp frontend and slapped it on a random computer
I hope his clients are aware
As long as they still have someone verifying and correcting mistakes, that’s a decent solution. It doesn’t need to be complicated. Whisper is by no means perfect, but from my experience it handles 90% fine, only having to listen and correct 10% is a huge improvement over transcribing everything manually.
The issues with so much of the current AI use is implementations that completely remove trained human oversight from the equation because the AI companies position it as a complete replacement for employees.
from my experience it handles 90% fine, only having to listen and correct 10% is a huge improvement over transcribing everything manually.
I’m curious how this is supposed to work? How do you only listen to 10%? How are you supposed to know which 10% to listen to? Wouldn’t you still need to listen to it all to find that 10%?
Yes, but listening to it all once and correcting the mistakes as you hear/see them would be faster than transcribing the entire thing, then proof reading.
All these articles, nobody posting error budgets or Delta’s with real doctors in this shit.
But we all know doctors are perfect with perfect notes every time and every nurse who gives you a dose 🤷♂️
God forbid that doctor take 2 minutes to confirm accuracy.
but it’s 90% accurate!
Ignoring that this means that one in every ten patients is getting inaccurate notes.
This will literally lead to people dying.
How accurate are the doctors? How accurate are they both in tandem? Are they higher than 90%?
Dual systems working on a decision could be significantly safer.
But Lemmy hates AI and loves to huff its own farts
It’s a fair point. Unpopular in this community, but fair. Of course, nuance is rarely rewarded in online discussion
“If you’re not with us, you’re against us. Huh. If you’re not an apple, you’re a banana.”
-Terry Pratchett, Thud
Fediverse has turned into the truth social of disenfranchised progressives. Don’t get me wrong I agree with many themes in this place, but the tribalism and unwillingness to refrain from immediately labeling a topic or person a total bad actor is pretty high. Everyone is a mustache twirling billionaire, redneck who hates trans and are Nazis, or a progressive hero.
I like to eat apples and bananas.
