@emilymbender

If you thought social media posts after death were horrifying, this is what is awaiting you beforehand.

Thanks for keeping an eye on these things!

@emilymbender
I have a feeling that this would be no more accurate to someone's wishes than if decisions were made by just flipping a coin.

The point is not to determine the person's wishes, but to take the decision out of human hands so that no one has to feel responsible for it. Just flip a damn coin, then, and call it a decision made by Fate.

@emilymbender Thanks for this very important article Pr. Bender! I shared it on my linkedin, I worked during 4 years in public health here in Montreal (for University of Montreal), and I strongly believe that your opinion has to be heard. This is quiet scary...

@emilymbender

Problem: "most critically ill patients have never engaged in comprehensive advance care planning"

Solution: invade patients' privacy, make them even more leery of bringing up sensitive topics with their providers, entrust lots of sensitive (and marketable) data to a black box system that we already know will produce biased output, and use it for life-and-death decisions

But there's not much point debating it because "it's inevitable"🤦‍♀️

@emilymbender No words. I can’t even…. Madness!

@emilymbender why not just prioritize getting patients to file their living wills/directives/authorizations instead of subjecting them to this appalling invasion of privacy and dehumanization? Because it’s not as sexy as AI that’s why.

@emilymbender 😳😮

@emilymbender yeah, let's have these decisions done by a glorified autocomplete function...

@emilymbender
The fact this was even proposed...

Welcome to The Future.

@emilymbender was this exact scenario a Twilight Zone, an Outer Limits or a Star Trek cautionary tale?

Let's go ahead and implement it anyway. We gotta do something with the trillion dollars of GPUs we bought.

@emilymbender (Yakov Smirnoff voice) In AI dystopia, computer unplugs YOU

@emilymbender

"We published this Viewpoint because it is very interesting, somewhat scary, and probably inevitable."

https://jamanetwork.com/journals/jamainternalmedicine/article-abstract/2821367

#EndOfLife
#PalliativeCare
#AI

@emilymbender "Why yes, if your husband Jim were able to speak, he would definitely want you to end his life as quickly as possible to free up this bed for paying customers. Have a great day!"

@emilymbender 👍 I believe medical science shouldn’t platform automatic care period.

@emilymbender

Making end of life decisions for a patient without adequate information terrifies people. Understandably. (It terrifies people who are primarily focused on not being sued in a special way.)

If you can avoid making the decision entirely? Hand it off to a (supposedly) impartial summarizer of The Truth (tm)?

What better way to avoid the existential dread and fear of being human in a difficult situation?

The last thing I want in life – literally? – is a chatbot deciding whether I live or die. https://buttondown.email/maiht3k/archive/medical-science-shouldnt-platform-automating-end/

@emilymbender my response upon reading your description of the article.

@emilymbender

as I read the article, the authors seem very well aware of the issues, and seem to be aware, as experts, fo many issues you do not seem to be aware of

@emilymbender
this is the editors note that goes with the article

read the last sentence, svp

@emilymbender mais non ???!!!

@emilymbender I gasped out loud when reading the title. And then I saw that they imagined that based on full audio recordings of doctor-patient conversations and I noped the hell out of there.

In 2005 I decided what I wanted to do professionally was to use computers and maths to improve medicine, which was super niche back then. How far I was from imagining the nightmare into which my dream would turn.

@emilymbender Getting ended by an AI deciding for you because of your meme search history is such a dystopian way to go!

@emilymbender @catsalad As someone who had to make those very decisions for my husband, who was unresponsive and dying from a glioblastoma: I am glad I have never met these ghoulish idiots in person, because my children need me to stay out of prison.

@emilymbender another one for @ianb, and quite a doozy, to boot!

@emilymbender If you could do this with a comatose person you could do it with a living person. They're proposing that a statistical model can be built to mimic telepathy to know the thoughts of another person.

If true, dystopic. But I don't think it's true, and definitely not with spicy autocomplete models.

@emilymbender
Ouch. And:

"must be grounded in an* understanding not only of the actual"

@emilymbender
Well of course it's inevitable that we'll all keep outsourcing our thinking to MOLE (Machine Only Learning Emulators). If we didn't, we'd only let our emotions cloud our decisions about letting clouds think for us. In fact, I think it makes sense to outsource the decisions about when to outsource our decisions too.

What could go wrong? : P

https://strypey.dreamwidth.org/1120.html

@emilymbender the level of sheer ignorance and incompetence required for this crap is literally "people should lose their licenses permanently" level.

See also: https://spectrum.ieee.org/how-ibm-watson-overpromised-and-underdelivered-on-ai-health-care

@emilymbender@dair-community.social truly horrifying. I'd personally be more comfortable with a coin flip.