Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
×
AI Medicine

Nurses Say Hospital Adoption of Half-Cooked 'AI' Is Reckless (techdirt.com) 106

An anonymous reader quotes a report from Techdirt: Last week, hundreds of nurses protested the implementation of sloppy AI into hospital systems in front of Kaiser Permanente. Their primary concern: that systems incapable of empathy are being integrated into an already dysfunctional sector without much thought toward patient care: "No computer, no AI can replace a human touch," said Amy Grewal, a registered nurse. "It cannot hold your loved one's hand. You cannot teach a computer how to have empathy."

There are certainly roles automation can play in easing strain on a sector full of burnout after COVID, particularly when it comes to administrative tasks. The concern, as with other industries dominated by executives with poor judgement, is that this is being used as a justification by for-profit hospital systems to cut corners further. From a National Nurses United blog post (spotted by 404 Media): "Nurses are not against scientific or technological advancement, but we will not accept algorithms replacing the expertise, experience, holistic, and hands-on approach we bring to patient care," they added.

Kaiser Permanente, for its part, insists it's simply leveraging "state-of-the-art tools and technologies that support our mission of providing high-quality, affordable health care to best meet our members' and patients' needs." The company claims its "Advance Alert" AI monitoring system -- which algorithmically analyzes patient data every hour -- has the potential to save upwards of 500 lives a year. The problem is that healthcare giants' primary obligation no longer appears to reside with patients, but with their financial results. And, that's even true in non-profit healthcare providers. That is seen in the form of cut corners, worse service, and an assault on already over-taxed labor via lower pay and higher workload (curiously, it never seems to impact outsized high-level executive compensation).

This discussion has been archived. No new comments can be posted.

Nurses Say Hospital Adoption of Half-Cooked 'AI' Is Reckless

Comments Filter:
  • by jhoegl ( 638955 ) on Thursday May 02, 2024 @05:18PM (#64443460)
    I think I saw a recent thing on AI and "ask forgiveness instead of permission", but these fools are messing with data that should be protected and I guarantee you isnt.
    • Re:Not AI (Score:4, Interesting)

      by phantomfive ( 622387 ) on Thursday May 02, 2024 @05:39PM (#64443514) Journal
      There are a bunch of startups in the last 5 years or so who are trying to collect data in hospitals to prevent deaths. It's not an advertising model (yet) because hospitals are willing to pay.

      One of the easiest methods of preventing death is by catching sepsis early, when it's still treatable, and a number of startups have been able to achieve death reductions in pilot tests. Of course, this means that nurses need to spend more time entering data, which is annoying.

      This is not AI, it's just data science, but they call it AI because "hype." Funding is easier, getting clients is easier, etc.
      • by narcc ( 412956 )

        This is not AI, it's just data science,

        What do you think AI is, exactly? AI is applied data science ... and good bit of applied data science is AI.

        AI is a broad field that encompasses a lot of things that you probably don't think of as AI such as linear regression, clustering, and decision trees. I understand this annoys some people, but I can assure you that this isn't anything new. This is what AI was always about. The science fiction version of the term came much later.

        Yes, there is a lot of hype around AI at the moment and the marketdroids

        • Re:Not AI (Score:4, Insightful)

          by Mr. Dollar Ton ( 5495648 ) on Friday May 03, 2024 @01:17AM (#64444030)

          LOL. Only 25-30 years ago calling linear regression (and a bunch of other well-established statistical techniques) "AI" would have you laughed at. These days... Sad, really.

          So, yeah, "AI" is "applied statistics", but with a caveat - it is "applied statistics" when the people who apply it know it, and know why and how to apply it.

          Just like in the olden days, when it wasn't adorned with meaningless buzzwords, this ain't generally the case. Instead, most "applications" of "AI" are firmly in the category "throw shit at the wall and see what sticks". With the addendum, that since it never sticks well, even "results" that represent the trace of shit sliding down the wall are deemed "acceptable science".

          • Re:Not AI (Score:4, Insightful)

            by serviscope_minor ( 664417 ) on Friday May 03, 2024 @03:56AM (#64444174) Journal

            LOL. Only 25-30 years ago calling linear regression (and a bunch of other well-established statistical techniques) "AI" would have you laughed at.

            It would not. I don't really know what to tell you other than that. I did my degree back in the late 90s/ early 2000s before the current AI hype/boom and comfortable after the previous one. And 25-30 years ago here's how it looked:

            When people aren't being nutty, AI isn't "THIS WILL REPLACE ALL HUMANS OMGLOLWTFBBQHAX" it's more like techniques you can use which would have otherwise required human intelligence. Note that does not imply general intelligence or even a particularly high level of it.

            So it was a pretty wide grab bag of topics. It included things like expert systems, logic systems, various algorithms (like A*, Dijkstra), and certainly included a bunch of ML topics. I can't remember which ones off hand because there was a political tug of war between two different departments (one of which included the Bayesian Neural Network group) over who got to claim what, but that's pretty irrelevant really all things considered. There was definitely a system on rules based AI (lightly targeted towards game AI systems).

            But basically once a topic got big enough in its own right it got moved out of the AI part of the "information systems" module into its own section, always leaving a miscellaneous grab bag for AI.

            Stuff sticks all the time. After it sticks people stop calling it AI.

            In fact that line was a pretty common one even 25 years ago when people were debating "yes but is it reaaaallly AI" for various topics.

            As a case in point, the Perception Machine was definitely considered an early attempt at AI. It was basically the first system for linear regression classification, and sure linear regression is now considered an established statistical/ML technique. But that kind of supports the point that people work on AI and once something becomes well established enough it gets its own name.

          • Re:Not AI (Score:4, Informative)

            by serviscope_minor ( 664417 ) on Friday May 03, 2024 @06:06AM (#64444298) Journal

            LOL. Only 25-30 years ago calling linear regression (and a bunch of other well-established statistical techniques) "AI" would have you laughed at.

            Following up: this is absolutely 100% demonstrably not true and should not be modded as insightful.

            AI: A Modern Approach (2nd ed.), page 720. Published 1995 and 2003. Clear from the preface to the second edition that linear regression was already in there.

            https://github.com/aAmitSengar... [github.com]

            I think a lot of the modern AI hype is stupid, but blatant historical revisionism doesn't sit well with me either.

            • What?

              • You said: 'Only 25-30 years ago calling linear regression (and a bunch of other well-established statistical techniques) "AI" would have you laughed at.'

                I provided a well known textbook from broadly that time period which has linear regression as one of the many AI topics the textbook is about.

                In other words your claim is wrong.

                • I have no idea what the book which title page you've posted says.

                  Linear regression is a topic in virtually every econometrics textbook. It ain't called "economic theory" on the inside.

                  Your point again?

                  • Looking at the index of your book, linear regression is mentioned in the expected context - as one statistical method of parameter estimation in chapter 20 of a book of 1400 pages.

                    I guess that pretty much settles it.

                    • linear regression is mentioned in the expected context - as one statistical method of parameter estimation

                      How to do statistical parameter estimation is a topic in a book about how to do AI.

                      I don't really see how this is so hard to understand.

                    • Yes, it is. It is the same topic in a book of physics and economics. It ain't "AI", it is statistics.

                      I don't really see how this is so hard to understand.

                    • by narcc ( 412956 )

                      Does it hurt being a stupid as you? That's what AI is*, you drooling moron.

                      Here's a link [towardsdatascience.com]. This is basic shit.

                      *This is an oversimplification, but you're obviously too stupid for anything more nuanced.

                    • Thank you for so brilliantly confirming my point, youngling :)

                    • by narcc ( 412956 )

                      Why am I not surprised that, in addition to being completely ignorant, you're also illiterate?

                      Here's a tip for you: When you're proven wrong over and over again, the correct response is to either admit that you were mistaken, or shut your stupid mouth out of embarrassment. Doubling-down just makes you look worse.

                      In your case, however, your best bet is just to stop posting entirely. You'll spare yourself the embarrassment and we won't have our time wasted with your nonsense.

                    • You were obviously given this advice many times.

                      You should follow it for your own good, instead of forwarding it needlessly to people ;)

                    • by narcc ( 412956 )

                      Wow, this is beyond pathetic. You were wrong. Get over it.

                    • The projection is strong in this one... And it is tireless :)))

          • by narcc ( 412956 )

            LOL. Only 25-30 years ago calling linear regression (and a bunch of other well-established statistical techniques) "AI" would have you laughed at.

            LOL! What's amazing is how confident you are, despite obviously having no actual knowledge or experience.

            You don't have a clue. Instead of posting laughably uninformed nonsense on Slashdot, try reading a book. Not only will we be spared your brain-damaged comments as you struggle to work out what those weird squiggles on the page mean, there's a very small chance that you'll actually learn something. If we're lucky, you'll at least realize how ignorant and stupid your comments are and stop posting out o

        • AI is a broad field that encompasses a lot of things that you probably don't think of as AI such as linear regression, clustering, and decision trees. I understand this annoys some people, but I can assure you that this isn't anything new. This is what AI was always about.

          You are describing "weak AI" and it's not what AI was always about.

          When people say "AI" by itself, they typically mean "strong AI," but not always.

          • by narcc ( 412956 )

            To the surprise of absolutely no one, you don't know what you're talking about.

            and it's not what AI was always about.

            That's simply not true. The term "AI" was coined by mathematician and computer science pioneer John McCarthy for the famous Dartmouth conference in 1956 and is exactly what I described. As the earliest use of the term, what I described is necessarily "what AI was always about".

            "Weak AI" is a term coined by philosopher John Searle in 1980. You'll notice that 1956 came before 1980.

            When people say "AI" by itself, they typically mean

            When the average moron says "AI" they typically m

      • What data do Nurses have to enter? If it's patient vital signs, this could come from existing meters, or something else, a new connected capture device... is "there an app for that?"

        I've seen Nurses do data capture in ICU, they have to enter data on a shared PC, not very sterile, a bottle neck and takes up space. Prefer to see them keeping an eye on patients.

        Interesting tech, but have they forgot about making it easy for their users?

      • In medicine, "AI" used to mean artificial insemination.

    • In this case it isnâ(TM)t really AI, it is ML on patient data. Basically someone implemented Prometheus with predict_linear on various metrics.

  • by Big Hairy Gorilla ( 9839972 ) on Thursday May 02, 2024 @05:23PM (#64443470)
    I know a hospital administrator, she's all in on convenience.. for the hospital administrators.
    Testing? Fit for purpose? Secondary considerations. Pass the Kool Aid.
  • Ai Nurse (Score:5, Funny)

    by labnet ( 457441 ) on Thursday May 02, 2024 @05:25PM (#64443478)

    I empathise with the nurses. It will be like Microsoft's level one help centre.

    You: I'm having an issue with my Arm. Whenever I try to lift it, it falls off.
    Support: Oh no, that sounds really frustrating! I can imagine how annoying it must be to have your work disrupted like that. But don't worry, I'm here to help and we'll get through this together.

    You: Okay, what should I try to fix it?
    Support: I totally understand your eagerness to resolve this quickly. Having arm issues can be such an inconvenience in our busy lives. But before we jump into troubleshooting, would you mind taking a deep breath with me? These kinds of problems can really raise our stress levels.

    You: ...No, I'd just like to know how to fix the arm falling off please.
    Support: You're absolutely right, let's focus on the task at hand. I don't want to make this any more difficult than it needs to be. I'm picturing how frustrated I'd feel if I were in your shoes. But please know I'm giving this my full care and attention. Have you tried rebooting your arm since the crashes began?

    • I swear I just got through a session with Esty chat support that sounds just like that: canned empathy. Real people did answer questions, but were aided by something making such fluffy statements in-between.

      It's like the AI version of that phone-wait message that says, "Your call is very important to us" while you queue for half an hour.

      I want deeds, NOT words!

      And it would be more entertaining to have a sarcastic Bender-style bot, as long as I can call it nasty names back. Chewing out a LoutBot makes the wa

    • "Elevation in heart rate detected. Would you like a pill?" -- Elysium
  • by sdinfoserv ( 1793266 ) on Thursday May 02, 2024 @05:32PM (#64443498)
    Even though health care groups claim to be "non profit", they continually reinvest in buying up other health care competitors, and huge payouts to C levels. Gregory Adams, Chairman and CEO kaiser permanente is paid $17,268,060 annually. The only way to pay someone $8,302 dollars per hour is replace real workers with automation. That's the goal of AI. That's how it's being sold. This is why the US spends more on healthcare per capita than any other country and has the worst results. greed.
    • Kaiser Permanente makes $17 million every 90 minutes, so it doesn't take AI replacing people to afford that salary. Just a $1 average increase to their members' annual premiums would almost double the CEO's salary.

      CEO salaries are simply too small to have a measurable effect on consumer prices or employee pay. There aren't enough highly paid executives compared to the number or employees or customers.

      • by Anonymous Coward on Friday May 03, 2024 @01:46AM (#64444052)

        There are certainly CEOs out there being paid insane amounts of money (like Musk's pay package) which absolutely do move the need on profitability and employee compensation. That said the issue with CEOs like Kaiser's being paid a lot isn't that the company can't afford it, it's that it's a sign of how top to bottom the company exists to waste money which drives up cost at the expense of patients and insurance payers (including to some extent taxpayers).

        If you've ever gotten health care in, say Japan, you'd immediately see how they don't wantonly waste money on everything from carpet to unnecessary gadgets and it's obvious why their health care costs are so much more reasonable.

        By the way the same thing is true of higher education, the community colleges around here have far nicer campuses than most national-level Japanese universities. And the tuition there is something like $5k a year. Walking around an American university campus it is mind-blowing how nice it is, and at the same time you know how totally unnecessary it is. Most classes can be taught with a chalkboard and a bunch of chairs, and professors don't need more than an ordinary computer to do their research (putting aside the physicists who need to smash atoms etc.). Yes libraries are needed too but the budgets for universities are far more than the cost of keeping the library current. It's a race to the top for amenities to attract students and faculty and it all comes at the expense of students, as well as faculty who turn into grant-writing machines.

        • by nightflameauto ( 6607976 ) on Friday May 03, 2024 @08:49AM (#64444558)

          There are certainly CEOs out there being paid insane amounts of money (like Musk's pay package) which absolutely do move the need on profitability and employee compensation. That said the issue with CEOs like Kaiser's being paid a lot isn't that the company can't afford it, it's that it's a sign of how top to bottom the company exists to waste money which drives up cost at the expense of patients and insurance payers (including to some extent taxpayers).

          If you've ever gotten health care in, say Japan, you'd immediately see how they don't wantonly waste money on everything from carpet to unnecessary gadgets and it's obvious why their health care costs are so much more reasonable.

          By the way the same thing is true of higher education, the community colleges around here have far nicer campuses than most national-level Japanese universities. And the tuition there is something like $5k a year. Walking around an American university campus it is mind-blowing how nice it is, and at the same time you know how totally unnecessary it is. Most classes can be taught with a chalkboard and a bunch of chairs, and professors don't need more than an ordinary computer to do their research (putting aside the physicists who need to smash atoms etc.). Yes libraries are needed too but the budgets for universities are far more than the cost of keeping the library current. It's a race to the top for amenities to attract students and faculty and it all comes at the expense of students, as well as faculty who turn into grant-writing machines.

          OLD MAN RANT WARNING:

          I'm just old enough to remember going into the hospital as a kid and it was a simple brick constructed building with offices (both public and non-public) and waiting rooms and that's pretty much it. Now going to the hospital is like walking into a cathedral. Waterfalls in the entryway, which is of course four stories tall to show the magnificence of the great healing facility, with murals dedicated to various healers over the centuries, a long stream for well-wishers to toss coins into, a huge cafe style eatery (which admittedly serves great food), the whole thing is a giant spectacle. And that's not even getting into how wildly they furnish the waiting areas now.

          I don't bother for myself, of course. I'm out of tossing tons of money at nothing that ever helped me. But my mother married an insurance salesman that got her hooked up for life. When he passed, her coverage switched to what is basically a free pass forever on everything so she's well taken care of. I've been running her to a lot of appointments as she goes through some back issues during her waning years. Every time I enter one of these testaments to greed I wonder how anybody involved in it thinks it's the right thing to do. Year after year we get told that medical expenses are going way up, and insurance premiums have to cover those expenses. Yet they have plenty of money to continually build out or buy up these giant, sprawling, luxurious, fanciful buildings when medicine should be about service to the people and utility. I'm not saying give us white rooms and lab coats only, but for fuck sake there's a long distance between what I saw as a kid and what I'm seeing now. And you can't tell me that shit has no effect on end-user premiums or costs.

          If we're going to continue to insist that our for-profit medical insurance system is the only way going forward? We need some form of regulation of the entire medical system beyond just, "You have to buy insurance, and they can decide whether you live or die. Fuck you." This "gimme more gimme more" thing is literally killing us.

          • Have a relative who worked hospitals nursing, nursing department head, administration, CFO from 1956-2010.

            - Medicare paid for just about anything form 1960s to 1980. Hospitals routinely billed just about anything they could to Medicare to maximize revenue. This caused a dramatic rise in hospital and medical industry costs
            - Insurance companies are regulated by the state government and, in many cases, have a fixed by law profit margin of X percent of premiums received. Insurance companies want a generally

      • Don't normalize this obscene greed. Gregory Adams pay is the 67th highest CEO pay in the nation. Meanwhile nurses go underpaid, med techs and cna's have to fight for wages barely livable. Emergency rooms are packed and people with life threatening issues have to wait hours. This is not normal. But it is a moral crime.
  • I know AI proponents claim AI is great and all, but is there any hard data that it actually improves anything?

    • Re:Does it work? (Score:5, Interesting)

      by timeOday ( 582209 ) on Thursday May 02, 2024 @05:44PM (#64443522)
      That would have to be addressed on an application-specific basis. Debating the issue "in general" doesn't make much sense.

      Definitely humans provide care in a way an algorithm cannot. On the other hand, when the summary does get around to mentioning what Kaiser is actually doing, it's monitoring sensors. Computers are pretty damn good at maintaining vigilance in sensor monitoring, in a way people are not. That's why we have fire alarms even though humans can also feel heat and smell smoke.

      • I'd definitely trust an AI more than a human to continuously monitor my vitals. At least until the hospital decides there's less need for nurses to visually check, and then starts firing nurses until they're up to their eyeballs in work again.

      • On the other hand, when the summary does get around to mentioning what Kaiser is actually doing, it's monitoring sensors.

        It's not doing that. From one of the links (https://about.kaiserpermanente.org/health-and-wellness/our-care/behind-the-scenes-alert-system-another-set-of-eyes?ref=404media.co):

        Every hour, the program automatically analyzes hospital patientsâ(TM) electronic health data. [...] the program uses a powerful analytical engine that takes into account many patient factors. These factors inc

    • by Tablizer ( 95088 )

      is there any hard data that it actually improves anything?

      Yes, ChatGPT clearly states that it's cheaper and better than humans.

      ChatTP [i.redd.it]

    • I know AI proponents claim AI is great and all, but is there any hard data that it actually improves anything?

      There's a few studies showing that certain specific tasks get faster with the application of what we currently call AI. And the decision makers, the investors (it sickens me that investors are holding all the cards in our medical facilities now), LOVE faster. Faster = cheaper = more earning potential, so go, go, go, and care be damned.

    • by dvice ( 6309704 )

      > I know AI proponents claim AI is great and all, but is there any hard data that it actually improves anything?

      Yes. AlphaFold had the largest impact, because they created a database of 200 million proteins with it. If you don't think that this is the biggest thing in the history of humans, you need to watch for example this video where they explain it to you:
      https://www.youtube.com/watch?... [youtube.com]

      Here is one example how the data was actually used, "the first successful application of AlphaFold to hit identifi

  • The problem is that healthcare giants' primary obligation no longer appears to reside with patients, but with their financial results

    Not yet, otherwise they would just let patients die without enough nurses, and without AI.

    • by Anonymous Coward

      The problem is that healthcare giants' primary obligation no longer appears to reside with patients, but with their financial results

      Not yet, otherwise they would just let patients die without enough nurses, and without AI.

      Private equity and some of the other healthcare giants balance cutting and overworking staff with keeping the number of deaths just below where states or families start to take action. If some technology gives even more financial savings in the near term, they'll use it.

    • Not yet, otherwise they would just let patients die without enough nurses, and without AI.

      A dead patient is no longer a source of revenue.

      Big Health has no interest in curing you but also doesn't benefit from killing you.

      Their ideal strategy is to drip-feed you the most expensive and ineffective medicine while keeping you in a vegetative state on life support.

      • A dead patient is no longer a source of revenue.

        But their estate still can be.

      • Re:Almost evil (Score:5, Informative)

        by narcc ( 412956 ) on Thursday May 02, 2024 @11:34PM (#64443978) Journal

        Big Health has no interest in curing you but also doesn't benefit from killing you.

        Are you sure about that?

        If your carcass is taking up time and space that could be spent on a more profitable body, they're better off with you gone. Cured or killed doesn't make any difference.

        If you count the "heath insurance" industry as being part of "big heath", there are countless people they'd love to see dead. They want healthy people on premium plans, not chronically sick people making endless claims.

        • If you count the "heath insurance" industry as being part of "big heath", there are countless people they'd love to see dead. They want healthy people on premium plans, not chronically sick people making endless claims.

          Not true. ACA/Obamacare limits health insurance profits to 15% of payouts. So the more they payout, the more they can jack up rates, and the higher their profits.

          The technical term for this is "perverse incentive."

  • The nurses are opposed to replacing human nurses with machines, but there doesn't seem to be any description of using any AI-based system to do that.

    If the nurses are against "half-cooked" or bad AI, then they join the rest of the world in that sentiment. The described system seems to be an algorithm to monitor data that then alerts human nurses to do something. Sounds like a good idea, and one that is incremental in functionality to existing systems. This potentially makes nursing a little less strenuou

    • Re: (Score:3, Insightful)

      by ceoyoyo ( 59147 )

      People love to conflate things, particularly when their only argument is emotional. "An AI can't hold your hand" sounds a lot better than "we're concerned that automatic monitoring software will mean 20% fewer nurses hanging out at the nursing station."

      A pediatrician I worked with was asking me what I thought was going to happen with AI and healthcare, and pulled the same "AI can't hold the hand of a sick child and tell them it's going to be all right" line. I asked her when the last time she held a kid's h

      • by Luckyo ( 1726890 )

        >The future of healthcare is a nurse practitioner with a computer.

        That's present in many developed nations. Nurses don't really "hold hands" outside tiny amount of outlier cases and TV. Their job is mostly monitoring and adjusting things like medical dosage based on prescribed plan and correcting minor problems. And these plans are almost universally on computers nowadays.

        And if you ever worked with an experienced nurse, you know that they are exceptionally emotionally detached people in a professional s

        • by ceoyoyo ( 59147 )

          In most developed nations nurses are not allowed to do much actual medicine without a physician's order. That's changing, and I expect it to change a lot more. Thus "a nurse practitioner with a computer." That combination is cheaper and is likely to be much more effective in the near future.

          Yes, I've worked with lots of nurses. They do indeed hold childrens' hands and tell them it will be all right.

          • by Luckyo ( 1726890 )

            I have no idea about your experience, but I spent a large amount of my formative years working in hospitals and hospices, because I wanted a different experience entirely. I have performed nurse's duties and nurse's assistant's duties.

            And considering that you obviously didn't comprehend a word I said, instead arguing against something I never said, I'm going to assume you had the same problem with nurses you worked with

      • by dvice ( 6309704 )

        I think the best part of AI is that it doesn't need to hold your hand, because it can just cure you.

        AI detects plenty of issues from eyes, heart beat monitor, breath, x-ray. For example it can detect from x-ray that you have a diabetes and you have not been taking your medicine. Combine enough data and you can do miracles with diagnostics, compared to humans that are correct about 20% of the time.

        • by ceoyoyo ( 59147 )

          An enormous amount of medicine is placebo effect. That's why it still works even when the actual treatment is completely random. Physicians are trained to project confidence because it improves outcomes. Up until relatively recent times it was about the only thing that improved outcomes.

          The combination of a human whose job it is to both project confidence and make you feel better, plus accurate diagnostics and treatment of the underlying problem will be unbeatable.

  • Specifically, the Lavender killing software
    • Idunno. Covering for hostile combatants gets you on all kinds of shit lists. No computers required.

      • by Rujiel ( 1632063 )

        "Covering for hostile combatants" is an even worse take than it was a month ago now that mass graves have been found around Al Shifa full of hospital workers amlnd patients (some still in operational garb) with their hands ziptied. Some were tortured.

        You don't get to pretend to have a moral high ground when you siege hospitals full of refugees and deny them electricity and medicine. You don't get to blame people suffering amputations without anesthesia on hamas.

  • AI can't do what I do, not in the way I do it. I have a certain touch. A robot cannot replace my touch. Ask the people who are suing me for sexual harassment.

  • Did they hold your hand? Did they show empathy?

    In a lot of hospitals, you're doing well to avoid having somebody come into your room at 3 am, flip on the lights, just to take out the trash. If you need sleep, you'd better not try to do it in a hospital. I'm not sure AI would do much worse.

    • by Anonymous Coward

      I'm going to let you in on a little secret: How you get treated at the hospital depends on how you treat everyone you interact with from the second they wheel you in. They will treat you the way you show them you want to be treated.

      You're an asshole. More than that, you're a stupid asshole. If that weren't enough you're also an arrogant stupid asshole.

      If you want to get treated better, don't be an arrogant asshole. There's not much you can do about the stupidity, but if you're not an arrogant asshole,

  • That then Medi-Tron 5000 is going to miss true emergencies? That it's going to flood the medical staff with false alarms? That it won't do either but will result in other kinds of extraneous assignments? Or off-the-books cleanup work?

    Or that it works roughly as advertised, which is bad because it cuts into potential union dues by increasing capacity without increasing headcount of administrative staff?

  • I'd actually be okay with this, depending on implementation, except for that devil common to all AI systems- liability.

    When something goes catastrophically wrong, who gets the blame and how do you correct AI's integration into a systems approach?

    Is it the software vendor with crap algorithms, trainers with a series of check-boxes, or CEOs who demanded such systems be in use?

    This is indicative of a more pervasive problem- decisions being made by those who will not bear the cost of those decisions.

    Well, excep

    • This is good! MORE companies to sue! or your family to sue after they kill you.

    • It's the same problem with driverless cars, who's responsible if it runs you over? The driver who wasn't driving, the company that sold the car, the company that programmed the AI decision making? In my mind it would make the most sense that the doctor treating you would have the liability in this case, the AI is a tool like anything else and if your doctor hurts you by using it that's their fault for improper tool usage.
  • At some point we'll invent an AI that can be bitchy and condescending, and slap around old people when nobody is looking. :)

    (Tongue firmly in cheek. I know good nurses.)

  • I really doubt any real nurse said "half-cooked." That term is some kind of Engrish. This was Kaiser Permanente,a mostly Californian (plus other west coast states and CO) system.

    I feel for the nurses. They were already plagued by software allegedly-to-help-them that made middle managers happy but got in the way. There's already too much non-patient time in nursing and that burns nurses out.

  • These nurses are right, but they do not know the half of it. Wait until they discover hallucinations.

    • We're pretty much at enshittification already.
      • by gweihir ( 88907 )

        You are not wrong. But at this time they still provide some service. I guess the next enshittification stage is that an "AI" tells you that you are fine and then charges you $10'000 for that.

        • True - I'm just grumpy because of some enshittification I'm dealing with right now. I believe that AI in medicine means the end of progress in medicine UNLESS AI can genuinely make leaps of intuitive discovery, which I doubt. For doing routine stuff, sure, but otherwise? Nah. I believe that western civilization peaked on 20 July 1969 and it's been in decline ever since, alas.
          • by gweihir ( 88907 )

            I believe that AI in medicine means the end of progress in medicine UNLESS AI can genuinely make leaps of intuitive discovery, which I doubt. For doing routine stuff, sure, but otherwise? Nah.

            That is pretty much what I get from following AI research for > 30 years now and trying LLMs out on some exams I gave to students. LLMs are pretty good at looking stuff up (although they sometimes invent stuff), but they have zero insight and zero understanding (they cannot even "understand" slightly non-standard descriptions of things), hence they cannot "discover" anything, except maybe when it is just looking through masses of data with entirely conventional means. That will bring a bit of progress, a

            • Some of the problem may be what how we define "Research." If you tell a student in k-12 to do "research" this means go to the library , use Google, whatever, and look up what other has said. When you tell them to try and demonstrate something, that's 'proof' or 'demonstration.' So perhaps people unconsciously think that medical research is searching through tomes of papers that have been written. I agree about mass data analysis, though as AI DOES make stuff up, things will need to be verified. What I get
  • The expression is "half-baked." Sounds like an AI nurse came up with this complaint, in which case, point very much taken.

  • I personally hate AI being hyped up at the moment. Personally had a lot of conversations with i.e. chatgpt and ... the thing is just stupid in a lot of cases. But if there is one thing AI is pretty good at, it is finding patterns in lots of numbers: like BP, temperature, ... and flagging an issue. it is easy to miss something as a human. AI is always focussed. Sure it cannot replace the care a nurse does, but it sure can assist there and make their jobs easier.
    Read a story where they used AI to analyse e
  • this is being used as a justification by for-profit hospital systems

    Well, there’s your problem.

  • Already have expert systems IN. When a patent stops breathing or has no heartbeat - an alarm goes off. Fail to urinate - again an alarm. The ONLY thing that saves money is either discharge - ie none, or a downgrade (out of IC). Some believe Kaiser is replacing real shrinks with AI generated voice, and charging the SAME fees. That would not be AI, that would be deceptive substitution.

The nice thing about standards is that there are so many of them to choose from. -- Andrew S. Tanenbaum

Working...