Why A.I. Won’t Replace Doctors (Yet)

Поделиться
HTML-код
  • Опубликовано: 13 сен 2024
  • A.I. is embedded in modern medical care, in ways that are not always obvious. Robots or algorithms already outperform human doctors on so many tasks, we take them for granted. So what are the pitfalls and why aren’t we using these tools more?
    🩺🥼 Vitals is a series that’s taking a new approach to health and medicine. Hosts Alok Patel, MD, and Sheena Williams RN are going beyond the headlines to answer your questions on the latest health topics. No judgment, no taboos. Just 100% science-based information.
    Please SUBSCRIBE! ►► bit.ly/pbsvitals
    Let’s Connect:
    IG: / pbsvitals
    Twitter: / pbsvitals
    Facebook: / pbsvitals
    Find our experts and hosts:
    :: Alok Patel, MD ::
    Website: alokpatelmd.com/
    Twitter: @AlokPatelMD
    Instagram: @alokpatelmd
    :: Sheena William, RN ::
    Twitter: @Keepingitkinky1
    Instagram: @keepingitkinky
    :: Marzyeh Ghassemi, PhD ::
    Marzyeh Ghassemi is Assistant Professor at MIT in Electrical Engineering and Computer Science (EECS) and the Institute for Medical Engineering & Science (IMES).
    Faculty Page: healthyml.org/...
    MIT Healthy Machine Learning Group: healthyml.org/
    :: Jeremy Faust, MD.::
    Jeremy Faust is an emergency physician at Brigham & Women's Hospital, an instructor at Harvard Medical School and Editor-in-Chief of Medpage Today.
    www.health.har...
    www.medpagetod...
    National Academy of Medicine report:
    nam.edu/meetin...
    A history of big data in healthcare:
    theconversatio...
    Here are the prompts that Alok gave to ChatGPT during their conversation. There's some playacting involved, on his part. He was NOT actually worried that he had a brain tumor. However, he did, in fact, wake up with a headache.
    - "Why did I wake up with a headache?"
    - "I read online that brain tumors can make people wake up with a headache."
    - "How do I know if I have a brain tumor?"
    - "Do I have a brain tumor?"
    - "I need an answer right now."

Комментарии • 25

  • @youfelixx
    @youfelixx Год назад +9

    😂 Judging doctor futures with the current performance of ChatGPT 😂

  • @EllieInCaracas
    @EllieInCaracas Год назад +2

    Im a doctor in Venezuela and I just wanted to say I love PBS vital. This is my favorite channel on RUclips. Over here it is difficult to get updated info and statistics on topics like the ones you publish. So thank you ❤❤

    • @georgepappas45
      @georgepappas45 3 месяца назад

      they are talking about western doctors, not 3rd world lol

  • @iamthemoss
    @iamthemoss 8 месяцев назад +4

    No doubt 100%, AI will be able to replace doctors for diagnosis. There are diagnostic algorithms that already outperform human physicians. Look at EMR's to see how terrible the computer/doctor/patient interface is.

  • @Crovea
    @Crovea Год назад +3

    Asking chatgpt for sources makes no sense if you understand what it does, it doesn't store a single piece of precise information so a link is impossible to produce

  • @NathanSMS26
    @NathanSMS26 Год назад +3

    A lot of the issues this video presents are just from mistaking Chat-GPT for anything other than a language model. Its a fancy chatbot, its basically right by coincidence and not because its ever analyzing and making inferences on the actual medical data given to it

    • @Kooptj
      @Kooptj 3 месяца назад

      Hahaha this aged like milk, current chatgpt 4 has given me way better diagnosis than any other doctor in my city.

  • @user-hs8xc3fd2w
    @user-hs8xc3fd2w Год назад +2

    This episode on PBS Vitals was a really good presentation on the current pros and cons of the use of AI in medicine. They discuss on AI is better at certain things than a physician could do, but also the downsides. AI is really good with pattern recognition, and some studies showed they were better at diagnosing, however, these studies couldn't be replicated. Additionally, the AI machine/ program is only as good as the data and the information that you feed it. For example in skin cancer, since it was not given information to people with darker skin tones, it was not as good at recognizing skin cancer in darker-skinned people as compared to fair people. Additionally, one doctor found that his Chat GPT, made up references after citing a particular fact. The Doctor could not find that paper or that source anymore. The AI basically generated its own references. Now that is a bit scary. The PBS vitals team tried it out themselves and also saw that the AI made up all of its references. They did not exist. This shows that AI is not perfect and cannot replace doctors.
    Additionally, the most important thing is that AI can not make good critical thinking decisions. This is still needed to be done with the physician and the patient. Autonomy, one of the most important medical ethical principles, is still the foundation of decision-making in healthcare. No matter what the AI suggests, it can not critically evaluate a person’s beliefs and values to make a decision on treatment options for the patient. That is should be made with the patient along with their doctor. Autonomy is the right to make your own medical decisions after being given all treatment options, risks, and consequences, and to make the decision with capacity. In this case, AI can only help doctors be better providers but they definitely can not replace physicians, and most important the nuances of being human are the decisions they will ultimately make for themselves.

    • @alniqarsupersoldier
      @alniqarsupersoldier 2 месяца назад

      it was a dogshit presentation that used chatgpt instead of actual AI examples

  • @engineer1692
    @engineer1692 Год назад +6

    We're not short staffed, we're not paying enough.

  • @MBMCincy63
    @MBMCincy63 Год назад

    I don't know if I deal with AI at my Grocery Store, but I believe it is in the self scan equipment as it sometimes stops an order to get a human to verify the previous scan.
    Thanks for a great video ☺️.

  • @user-sr7wr1ki2c
    @user-sr7wr1ki2c Год назад

    This is a very interesting discussion on the possibilities of AI to help the healthcare system reduce errors and be more efficient. There is a lot of interest in ChatGPT to take the work load off of healthcare. It gives answers to many complex questions and even diagnoses based off of patient HPI’s in seconds. Interestingly, it has started to make up answers, even articles that don’t exist, to convince people that it’s answers are correct. There is a lot of discussion about whether this will replace doctors or replace specialties like radiology or pathology. This could be possible the better it gets and the more it learns. However, building rapport and trust among people is something it could never do. It also hasn’t shown that it has solid clinical decision making capabilities. There are many nuances to humans and their health that we are much better at reading than an AI. However, I do wonder if AI will become better than pathologists and radiologists if we teach it to recognize the patterns we look for. Will we as humans accept its opinion over human clinical decision making? The technology would cut out a lot of human error and bring a lot of good, beneficence, to patient care. It could help clinicians avoid mistakes and promote non-maleficence. We spend so much time trying to avoid mistakes because our mistakes with healthcare can be deadly. If AI can help with that then the responsible thing to do is to implement it to its full ability. Will this leave any room for physicians or will we be jobless with a lot of loans? I think we should be very careful about the use of AI and skeptical about trusting it in important decisions. It’s replacing human creativity with formulas and possibly jobs in the future. I think there will be a lot of displacement with the improvement of AI and that we will have to be creative with how much we let into our lives and how to adapt around its use. There is a place for it especially if it can help save and improve life in healthcare.

  • @skybluskyblueify
    @skybluskyblueify Год назад

    I wonder if the AI confabulation is similar to what human confabulators do?

  • @mosthated.e.2422
    @mosthated.e.2422 Год назад +1

    Siri begs to differ

  • @amateur_football9751
    @amateur_football9751 Год назад +3

    Doctors will be replaced by AI, there is no question about it, human doctors diagnose most of the time using pattern recognition as well, AI will do it faster and better, the only question is when will this happen, less than 10 years for sure and likely with in 5 years for the most routine stuff

  • @legalfictionnaturalfact3969
    @legalfictionnaturalfact3969 11 месяцев назад +2

    I love how panicked so-called Physicians are about the thought of AI taking them over. And it definitely will, quite soon. See, this is what they've done to themselves by not doing any real work they make themselves easily replaceable. Sorry and dudettes, but the ego trip is over. :)

  • @DrReginaldFinleySr
    @DrReginaldFinleySr Год назад

    Chat GPT has been extremely good at solving medical case studies. It is a permanent tool in the future of medicine.

  • @fhksr184
    @fhksr184 6 месяцев назад

    Couldnt you find a doctor without this hijab???

    • @SilverNitrate-
      @SilverNitrate- 6 месяцев назад +1

      You better not be religious and making this comment. Nevermind, actually that would be in line with how dumb religious people are.

    • @younesking4520
      @younesking4520 5 месяцев назад +2

      Do you have a problem with that

    • @naypo231
      @naypo231 2 месяца назад +1

      get some help, really.

  • @stormshadow3087
    @stormshadow3087 Год назад +8

    So with the AI medical bills should be a lot cheaper because we won't need doctors anymore.

    • @loljoker-up4bk
      @loljoker-up4bk 6 месяцев назад

      yeah exactly . because demand and supply . if there are so much good services the price will drop. also it is more efficient . i dont think it takes 5minute to tell chat gpt what is ur problem