[Music]
hello and welcome to technically speaking where scientists and Engineers come together to chat about a common interest to share knowledge and satisfy some curiosity I'm Laura and I'm joined by Sarah and Antonia to talk about artificial intelligence and how it could be used in healthcare to help with Diagnostics or what else it might do for medicine so Antonia I think the inspiration for this episode started off with you so tell us about it so I was reading an article and the headline said
chatgpt4 can pass the US GP medical exam and I thought hmm interesting we've heard all sorts of things coming out of chat GPT and I thought about one of my friends has recently passed her GP exam in the UK maybe it'd be a good time to talk about it fair enough and that's you Sarah so you're a medical professional so um do you have any thoughts about this just a caveat I actually haven't passed my GP exam yet I've passed my GP entrance exam so as the exam to get into
GB training so I'm both a medical professional and a surgical educator my initial response to AI or the concept of AI passing these exams is it it threatens my job um so that's my immediate gut reaction and I also know that it's going to be heavily invested in because the NHS will need to reduce Workforce Workforce costs over time so I think that it's going definitely going to be an area of interest in the future all right and I guess one big question there
is is it better to have an AI doctor or GP or surgeon even or is it better to have a human doing that job and what's the difference so I guess we should start as we always do in this show with defining what AI is now I know it's something that needs to be trained by feeding it data I don't really have much experience that I've never used GPT I've heard a lot about it but I don't know how any of these things actually work so if you guys got an experience of AI or have a better
definition I don't have a definition as much as just a AI stands for artificial intelligence and I think it it gets mixed with some of the terminology like machine learning and neural networks and I think those are two of the key tools that are making up the AI that we see around at the moment such as in chat GPT or Bing's search engine is now also AI powered and my experience is we've got a bunch of data and fed it into the machine as it were and told it to either
identify patterns or replicate something and try to do the next iteration that follows based on that information that's been given and so it's learning and sometimes it will develop a new way to pick up those patterns and so it becomes its own teacher and that we're not telling it exactly what pattern it should be reading the neural network I think helps it combine more things so it's not just a set data set that's my understanding as someone who isn't in computer science
I think I mean I'm also not in computer science but I think that the term artificial intelligence is a bit of a misnomer because it's only as intelligent as the information that we feed it so it is based on the data that we give it it can't create its own new data and it only learns as much as we reinforce that learning it it can't reinforce its own learning if that makes sense or if it does it creates its own feedback loop and in which case it has its own biases that will get Amplified
as it goes through more and more iterations in a popular culture example I want to say it was Twitter that tried an AI on people's tweets and then it found it got more aggressive more races more bigoted that they had to shut it down within a very short time frame because they realized that what people put on the internet should not be repeated and Amplified this is something that's I think is important to consider when we talk about bias in healthcare and bias and AI because any data that we
put into it any bias that we put into it will be perpetuated and there's already something that I'm very passionate about is equity in healthcare and any inequities that currently exist in healthcare is only going to be perpetuated by this AI it can't go against any of the societal prejudice is that we have one of the things that's important to me is critical thinking so the ability to take in information and weigh it against other information and say well does that match what I already
know about a particular thing or am I spotting some sort of patent or something doesn't fit into that pattern that doesn't make sense that I should be questioning I I don't know if AI has that same capability to think critically about something and decide whether something is true or correct or appropriate or not I don't know is the honest answer I think it's able to say this doesn't fit the patterns I've been fed but able to truly assimilate knowledge and create new theories I don't think so
I think it'd be interesting because um like you know asimov's Laura robotics we put some certain rules in to ensure you know the Safety and Security of our future as we developed robotics at the time I'm not going to repeat them because I don't remember them off the top of my head but I wonder if we do put those barriers in can we put those barriers in in a non-biased way but also isn't too restrictive because then there's that idea that you've put in a limitation but
then does that create other consequences that we didn't foresee when we release it out into world and actually it has a complete blind spot because we said you're not allowed to think about X so then never thinks about X I guess it depends what you're using it for I guess pop culture example is iRobot this robot was sort of almost like a human being and seemed to be doing certain things and seem really independent but a more practical example of what artificial
intelligence can be used for is I think I saw um something a while ago now I've mentioned it in a previous episode about training AI to be able to tell when a tumor is cancer or not based on an image of it and it was actually better at doing it than the specialist who had been using these images to try and diagnose cancer for decades maybe and it seemed like the AI had been fed sufficient data that it could spot I guess really subtle color using the images that sometimes a person might
miss because they're tired or they've not eaten or they're just having a really bad day or they're distracted by something AI doesn't forget and it doesn't get distracted necessarily so I can I can see how in that instance it's a lot more useful than just doing some big picture out in the real world doing everything thing yeah so for focused tasks absolutely and that specifically I was talking to some histopathologists um some colleagues of mine who very that to explain the term
um basically they're the people when you take a tumor out you send it off to the lab they're the lab people that then slice it up and look at it under a microscope and tell you what is in that biopsy um so I was speaking to some of them and they were saying that actually the way that they study for their exams is purely by looking at hundreds and hundreds of slides hundreds and hundreds of cases and that really interested me because actually that's exactly what we
do for AI is that we just feed it all of this data and eventually it comes to recognize those patterns the other thing that I was talking to them about because I raised the idea of oh well that's exactly what AI does so how is your job any different um and they were talking about actually it's something that they are worried about in the the world of histopathology that they now need to look at career tracks that are AI proof because there is AI in some hospitals in America and
they're about to bring it into a hospital in Nottingham as well to start looking at the more simple cancer slides add to free up histopathologist to do something else it's not a particularly attractive like career or part of medicine um which is possibly why they're so heavily investing in AI but also they're heavily investing in AI because it's so easily replicated by AI yeah and I imagine there are a lot more Niche things that those professionals could be doing not just looking at
images and saying yes cancer neural cancer they can be talking to people but actually so these specific doctors don't have any patient contact right what do they do genuinely as they look at slides and they tell you what is on the slides that is it that is their entire job okay so their job is effectively being taken by the artificial intelligence yeah which is why they've been talking about oh well you know if you go into things like prostate core biopsies which are much more
specialist and require a human eye as opposed to ai ai isn't as good good with those um that's like the route that they're going because they're they're already having to think about AI proof sub-specialities all right wow you get this idea that all medical professionals talk to patients all the time but I can see that's not necessarily true there are a lot of people in the lab just doing analysis supporting the people that are I guess on the front line it's one way of looking at it
yeah the the patient client facing part of a person those doctors did they go through the same training that you would have gone through so the same amount of like time and sweat to get to that point so we all have the same degree and then after two years of like basic training um and rotating through hospitals you then choose what specialty you want to do so at that point they they will have had a different life compared to me in that sense that means loads more trainee
doctors could be free to go to other disciplines it's just the challenge of people who are already trained in that field will have to find another field which they can apply their skills to yeah and um in a Utopia uh Ai and machines taking our work would mean that we have more free time and the ability to enjoy our lives but of course we understand that that's not how Society works so it is a threat to people's livelihoods and that I think is going to be one of the core
reasons why if you you speak to a doctor their initial response will be no don't like it this is my job only I could ever do this and AI of course is always going to be inferior but actually there will be some ways in which is superior I I can see again that um being able to diagnose something that frees up time to do something else would be particularly useful if you've got any specific examples of how it could be superior one of the main ways that I I think it will
be very useful is when you have for example healthcare workers who are working in isolation in remote remote areas so you're the only doctor for a community um it's useful to be able to bounce ideas off of something that we do quite a lot in hospitals as we speak to consultant colleagues about I've got this really interesting case and I'm not really sure what to do I think it might be this I think it might be that where should I go with this and you can kind
of have that conversation with an expert opinion um so I think it would definitely you know benefit people who are working in isolation um it might also improve access to health care so for example if you've got an impoverished Community then AI is going to be cheaper than hiring a doctor so it could also improve access to Healthcare in those sorts of ways statistician compared to an adequately trained doctor but you have to train those doctors so it does then mean that you don't have to
train people in that skill anymore um so you can cut out quite a lot of money and effort there if you wanted to I would argue that becoming overly reliant on AI could be quite dangerous though if for I mean stuff that happens all the time in the NHS like our servers go down and then we suddenly have no access genuinely we suddenly have no access to things like drug charts to things like um patients observations so like their blood pressure and their heart rate so
suddenly we have no access to any of this sort of stuff and we have to go back to being on paper so if we become overly reliant on AI we have to have the infrastructure there to support it all right you touched on a lot of points there one that I wanted to pick up on was the idea that yeah it'd be great to to have a sort of colleague inverted commas um as an AI to to discuss ideas but we've also seen like people talk about AI hallucinating and just getting basic
facts wrong that's what people have called it in the industry of just it just literally says so you could look up on a spec sheet sorry specification and it says this camera phone has this megapixels and instead the the AI just picks up a random one and just says now it's got like 50 billion I have heard this for teaching purposes essentially a lot of lecturers were saying that students are using AI to do the coursework for them and the air the eye had been um set up in a way that it
was deliberately putting errors in there so the lecturers could spot it so I wonder if that's what the Hallucination is about I I'm gonna let you into a little trade secret here the majority of my medical degree was learned on Wikipedia one of the one of the skills of being a doctor is not necessarily being able to have the entirety of medical knowledge in your mind but being able to to like to know where to go to look it up and you have to retain that skill even when
you're using AI if it outputs something that you go well that's a bit weird you need to have other options to be able to to know that it's you know giving you the correct answer I would agree with that actually because um I've been a scientist working in labs and supervising students for years now we would all say the same thing that it's not that you've got an entire textbook in your head you definitely have a lot of knowledge in there and that knowledge helps you infer things from other bits
of knowledge that you may be not quite as familiar with so same sort of thing yeah absolutely there was something you said fairly near the start about a patient facing roles essentially so I have a story from many many years ago I went to my GP my general practitioner saying I've got a sore throat and I've had sore throat for a while but I've not been able to do anything about it because of my situation my situation has now changed so I can get to you during
normal working hours take a look at it tell me what's going wrong you can definitely see something is wrong back there and they literally laughed at me an AI wouldn't have done that no it would have asked me questions yeah and Antonio you actually had a paper that you sent me today saying that AI has been proven to have better bedside manner um than doctors and I can entirely see that uh as you said we all are human we all have human error and empathy is proven to be a skill communication has
proven to be a skill it can be learned but people have to put the effort in to learn it so yes on the one hand AI would never make a mistake such as outputting hahaha are you giving them uh symptoms and I should joined it too unless you joined it too unless it's taught or that all of our imperfect data but I also think as I said it may be an equal or superior diagnostician but it might will not necessarily be an equal or Superior clinician and the reason why
I say that is because a lot of my job is delivering bad news especially diagnoses of cancer and for that you need a very specific amount of tact and the ability to respond to the patient in front of you to the person in front of you and yes probably eventually you could get to the point where if they look down they're sad or if they make that very specific twitch of their facial muscles they want to hug eventually the AI will get to that point but will will it ever be able to replace
the warmth and the emotion that comes from a real person connecting with you and holding your hand in that moment I don't think so no I think that human connection is what would make a big difference in a situation like that and I will say I've been treated by some healthcare professionals who have been absolutely amazing you don't have to defend them it's absolutely I'm just thinking of some nurse practitioners that are just not knowing exactly what to do and sorted things out
for me and just understood what I'm going through without me having to say anything to them other than my arm is swollen and numb help yeah yeah and I I would argue that it's their life experience that means that they're able to provide you with the exact support that you need and it's their experience of dealing with people who are similar to you and there's no reason why we can't input that data into an AI but do we want to yeah laughs is that what we expect you know
there's also that kind of managing expectations that you know someone's always wanting to go I want a second opinion I want the person with the most credentials or do we then start putting AI with the best success rate in front of them and say well they've got a 99.9999 correct rate true which you will never be able to say of a human absolutely I don't know thinking about the film AI the one with Jude Law and the other child actor that I can't remember the name of um
Haley Haley Osmond Haley Joel Osmond that was it anyway AI that film because they have robotic sex workers they had lots of companion robots who were able to emulate and then they had AI robot children that were able to emulate those sorts of human connections and human emotions um so I think the fact that we are creating films and literature about it probably means that actually ultimately we will end up wanting it um if you could and it's it's something I don't know that I've done with my
friends of like if you could build your perfect man what would they be like I mean if you could build your perfect doctor what would they be like and if you could why wouldn't you because they would be perfect yeah what I don't really know what I'd want a doctor to look like I kind of only go to them when I'm told to really like um oh well I've got a an implant that needs to be replaced every three years and I get letters saying you have to go for this test it's not often I get ill or think I've
got something that says I've got symptoms I should go to visit my GP or sit in an emergency room so I don't know I've not had a lot of experience on that front to be able to say what I'd want them to look like well like most of what I see is what I see on TV that's probably not very representative Gray's Anatomy yeah Antonio yeah I've I have been Cinder docs for about all sorts and also just when I had a niggle you know when I wasn't really sure what what it is but it had some sort of measure
measurable impact on my life like you know can't sleep as well because of said sad thing um and they do take a while to to get to the bottom of it you know there's a bit of trial and error that because you know it it's not always obvious what it is and I think I think part of it was drawing covert as well where we didn't have face-to-face interaction and so there was a little bit of like well we had um asked my GP so you could send photos which felt really weird like here
here's this weird patch of my elbow can you please see what it is I know black I can't really see anything it doesn't look that bad and I'm trying to describe it with my vague non-common medical terms because I feel like you know doctors have that like measure of is it a stabbing pain is it a throbbing pain is it a sharp pain and I'm like I don't know I've never been stabbed yeah this is my own experience of pain so yeah it's almost like you need a doctor to be able to interpret your
human foibles because like if an AI is only very good at understanding very precise instruction and we are imprecise with our uh descriptions then they would have to figure it out you know but I guess the argument is that over time with enough data put in they would learn that you know the majority of people when they say this they mean this and if it's unclear then they ask clarifying questions I can see that getting really frustrating though it's but like on that
the point Antonio made about pain if you're sickness saying well it's it's kind of starving but I don't know I also throbbed as well and it's like which is it is it stabbing or throbbing pick one but yeah and I keep I feel like I'm defending AI a lot but then you would train it to say okay that's it let's move on from that question because that's exactly what a doctor would do is they would just say you're getting frustrated by this let's move on let's just let's talk about something else
it's an interesting point though can you train AI to be I was going to say more human than a doctor with no empathy actually that's what I want to say though I mean I I remember reading somewhere but it might be apocryphal so please forgive me um but surgeons have a higher rate of psychopaths than any other profession um because they like the stabbing get theme now um potentially I mean we teach set phrases to medical students on how to respond to uncomfortable
outpourings from patients uncomfortable for the patients I'm uncomfortable for them so you say things like that must be really hard or I can see why that would be difficult for you or yes I I get why that's frustrating like you you teach them very set phrases but if you don't apply them correctly with the correct like vocal tone and with the correct facial expression it does come out like AI is just giving you yes I can see why that's hard oh that did sound a bit patronizing
talking about training them and pouring information into them you mean the AI not the DNA doctors yeah no not the junior doctors no um trading Ai and trading and pouring data into AI how much data is enough how much data is too much how much of our private lives should we be putting into these machines how secure is it um who has access to it who has the right to read it um I mean and at what point are we allowed to just not know stuff about ourselves like if you for example
if we put people's DNA and family history and medical history and everything else in that at what point it might come up saying oh you're at risk of x y and z do they have to know that do I have to know that oh what if it knows it and tries to stay towards it because you don't want to know yeah yeah that's a very cyclical argument yeah I guess a good example of what you're saying is there's a TV program it's on Netflix I think called The Bold type it's about three youngish women making
their way in publishing in New York um and one of them says um I think my mother might have had breast cancer and she ended up getting herself checked to see if she had the gene and she did and she got a bit paranoid about it it was doing all these checks and whatever else and eventually decided to get a mastectomy so she didn't have to worry about it and she thought that would solve the problem but she obviously got implants to replace what had been removed and then she started to feel
like it wasn't her own body and she was uncomfortable and there was nothing she could do about it and I think if I were in that situation I wouldn't want to know I'd rather take my chances because just because she have a gene doesn't mean you will get cancer right there are other factors at play but why if it becomes so good at predicting that it just knew because it's had so much information that the chance of it being wrong was so little if it could say with like 90 certainty that given your
environment your habits your genetics and all this other information that's not directly relevant to me well I have to sit there and think well is it if I am going to get cancer and I can avoid it maybe I should maybe I wouldn't want to I don't know maybe I'd be happy with waiting for that point and then having removed when the time came I don't know I suppose it's on the one hand you have a right to know and make an informed Choice with your own body on the other
hand it's the right to be ignorant because ignorance is bliss and if you know that you're likely to die of a heart attack at 75 that's so far away but it's always going to be hanging over your head if you know now know that um compared to if you were just blissfully up until the age of 75 living your life and then you know one day you just don't wake up on a lighter note I'm sorry I was going to continue down that model in veins this kind of thought process was applied
in the Black Mirror episode hang the DJ which was where a couple had an app and they got a number and it turned out that number represented how long the app expected them to be together so if you knew the number right from when you met how would you play out that relationship would it change the way you treated that relationship absolutely absolutely if if an app told me or you'll be together five years and then break up I'd be like no I'm just not gonna bother
it's been five years in a relationship that's gonna end see I'd go the other way and I'd want to know why it was saying that and try and beat that number I said you're not telling me it's gonna last like this it's going to be different it has to be different I refuse to believe that's true but and maybe maybe on on the side of things is that a sign that there are some people like me who will put their entire faith in what the computer tells them when it could be wrong
and are we you know are there people out there who will say oh I'm gonna die age 50 of Ms no I'm gonna kill myself at 40.
so I don't have to live that last 10 years and what if it's wrong you know how how much Faith are we going to end up putting in Ai and that is that is a danger that we shouldn't believe that computers are perfect AI is perfect because ultimately it is based on human fallibility so I guess you'd want some way of also training it to see when it's been fed something based on something that's a fallacy I really don't know how you do that though I I mean I guess you
could always just you'd do it with the caveat of like the same way that people oh that's just my opinion though of you know you have a caveat of this is AI and we cannot promise that it's perfect or whatever and in the same way that you know whenever I'm counseling patients I always say based on the information that I have this is what I think is going on therefore this is what I would recommend and there will be risks with everything and it might not turn out like this you
might end up having this problem you know doing thorough counseling on the risks and benefits of AI opinions and are they better than human opinions or are they still just opinions I guess from my side I always see my opinion as just that just my opinion this is my professional opinion but do patients ever see the advice the doctors give them as an opinion do they see it as gospel that is true I guess it depends on your experience with them I think some some
people just straight up don't listen if it's not what they wanted to hear right like you know you've been told you should cut this out of your diet because it's increasing your risk and people don't like people know doctors are right but ultimately they don't follow it ultimately they want to eat that lambdish even though it's going to make their gout worse yes yes this friend who spent a long time trying to get this swollen ankle diagnosed and it makes me wonder if they had AI well
the doctor had AI would it have spotted it faster because it was like sooner than the typical time that you would get this condition and also he didn't like going to the doctor because he couldn't walk so you just go ah it's a bit awkward now and then you know you know when you you go to a doctor and you say oh I need to see a doctor about this ah well the next appointment's like two weeks away and by the time that two-week appointment comes up you're like it's
gone but it comes back and so you'd book another appointment and then it's gone again and so they never actually get to observe it yeah um that's a really interesting point in the this friend of ours who's been diagnosed with gout he is very young to have gotten gout he is not the typical sort of person to get gout either so usually you associate a gout with older white men generally who drink a lot eat a lot of meat like think like old-timey ships captains that's
generally Henry VII or Henry VIII yeah exactly and that's generally who gets gout and he doesn't fit any of those stereotypes so would a computer that's purely basing it on pattern recognition ever come to that diagnosis um would it come to it quicker based on some unknown variables that we that we haven't inputted yet so the histopathologist I was talking to they were saying that actually the AI now uses criteria that they that they don't know it's using for it it's hard to
describe but basically they inputted the criteria that they've teached that they've taught trainees and it's also using some extra criteria that it can't explain that it's using so would eventually AI get to the point where it would come to these conclusions quicker because it has picked up extra patterns that we don't see as clinicians but then could that ever be the absolute truth because you couldn't explain it well I'm just wondering if this goes back to what
Sarah was saying about you wouldn't just rely on the AI alone for various reasons so if you can get it to explore blame to you so it can teach you what it's looking out for yeah because I guess that's something with people like well that looks like cancer to me but I can't explain why that's cancer I just know that it is trust me this is my opinion yeah trust me I'm a doctor I'm imagining this 200 300 years in the future is what I'm imagining where you have ai practitioners who are operating
independently and at that point I do wonder whether it would come to these sorts of realizations sooner or whether I think definitely in the short term it probably would never have considered it as a differential because it doesn't fit the standard patterns I guess there are always going to be outliers in the data because our understanding of the human body and well the entire world isn't perfect yet anyway there are still lots of things that we don't know and there
are still lots of extra patterns to spot and I guess that's also where some of the biases come in healthcare already like for example um women's problems are often underlooked people of ethnic minorities in a white country have particular likely outcomes that white people don't get um and that again it goes back to what you're saying about training it to already have bias because the data set is limited yeah absolutely to end on a slightly lighter and more futuristic you know I've heard
little bits about advances in surgery and like doing things that you would never have thought possibly before like using little robotic arms to do things and microscopic images so you can see things with Incredible detail is that something that you combine with AI so you've basically got a robot doing surgery on people with no human intervention so I spoke to a couple of my colleagues so I spoke to one person who is the head of Robotics at Leicester um University Hospitals of Leicester
um and he was saying that the main issue for humans is there's no haptic feedback during robotic surgery because you're essentially it's it's like piloting um a remote control car there's no like you can't feel when the car goes over a bump or you know Turns Upside Down you just have to see it and then base it on what you're seeing this is why I don't like computer games no feedback at that time I obviously need the haptic feedback and not just the visual yeah whereas with laparoscopic at
least you get some haptic feedback because you're directly touching stuff so sorry for laparoscopic it's Keyhole surgery so using a camera and looking inside using very small holes but you're still touching things directly whereas robotic it's a computer attached to that equipment and then you're you know I don't know a couple of meters away controlling that robot with like a little like joystick and two two joysticks genuinely two joist joysticks that move in three directions that you
can control all of the different um instruments with so you've got no idea if you're touching something that's quite squishy yeah and it's all based on your prior knowledge of what uh what anatomical structures are where which you could easily program into AI or you could map it you know in in this futuristic world you don't even need to give it an estimation you could just scan them true very true yeah you could do an MRI and then you could program that into a robot and away it goes
that's very true actually in which case if you were to do that that would actually overcome quite a lot of the stuff quite a lot of the arguments against um Ai and robotic surgery because the main problem is having the confidence to cut something when you're unsure what it is so when you lift something up for example and you go well I know it can't be this I know it can't be that and this other really important structure I know it can't be any of that so it can't it's
nothing important so I can just cut it and that is a bit of confidence the idea of non-important things in your body it's a bit but yeah like the appendix yeah but the appendix wouldn't be there but yeah let's just cut it out fans don't need it sorry anyway so it's having that confidence to know that there aren't any vital structures that you're going to go through basically and would and AI ever be able to independently make those sorts of assumptions and I suppose eventually you
would get the technology to the point where you could do an MRI scan but you would have to have a surgeon label every single structure as cut hair don't cut hair important not important because I mean it's even been it's been I mean anecdotally found that nurse practitioners nurse surgeons nose operators don't have the the confidence to cut without having a surgeon there saying yes to take it and having a person there to take that risk for them I feel like AI probably wouldn't do
risk very well yeah I can see why you know people have that kind of measure of how much risk they're willing to take whereas an AI we kind of say like you know if it was a programmer they might say okay let's say 99.99999 absolute but then do we get in a in a world where these robotic surgeons go into the operating theater and then go yeah that's a little that's a little bit too borderline for me yeah exactly everyone get back out we're coming out this yeah and I mean risk in surgery is
is I mean basically the whole premise of surgery is risk management so what what is the risk if we don't do the operation what is the risk if we do what is the risk if we do this specific operation versus a different type versus if we just do this or we just do that um so I I think risk management is going to be a big area that AI will struggle with um but then there's other stuff that's taking out uh surgery as a required specialty which obviously we won't go into here uh future episodes maybe
especially the the idea about risk which is the very first episode talking about what risk means to us do it over two years ago now yeah and we didn't go into that much depth though wouldn't AI be able to judge risk um to sum up my my opinion on this whole AI thing um AI in healthcare AI is built and trained on imperfect data from imperfect humans therefore expecting Perfection from AI is impossible we shouldn't expect Perfection from Ai and there is no such thing as
perfection in the human world I also don't expect Perfection from your doctors well that just reminds me of the Chaos Theory episode where we said you know with enough variables and enough computing power we could almost put a theory to predict something like this but will we ever get to that point where we have enough computing power to do that good question what do you recommend for people trying to train AI to be a better doctor because you know there are
pros and cons to it we've talked about a lot of con but they're they seem to be if we could work it out you know we're very early stage in it so what do you think should be taken into consideration today for the future what would be the ideal clinician and then from that build the AI into that so you don't just think about a diagnostician because we're basically there already with AI you also have to think about personalized management plans you also have to think
about communication with patients you also have to think about the ai's own resilience and ability to deal with the workload you have to think about the way that it interacts with other healthcare workers and the system in general that is probably what I'd recommend is think about what your ideal is and build the AI towards that in a holistic way so like any good project management yeah exactly great Okay so we've covered a lot of different things about Ai and how it could impact
Healthcare today or how it already impacts Healthcare thank you Sarah for sharing your experiences you're welcome I hope we continue to have great conversations like this in other episodes the views expressed in this podcast belong entirely to the person that said them they do not represent any industry or organization if you enjoyed listening to these views it would really help us out if you could rate US leave a review and tell a friend this podcast was sponsored by no one but if you're
interested in funding us to continue to have Frank discussions about science and engineering please get in touch [Music]