Surgeons of the future – AI

Group 22

We are living in the digital world where artificial intelligent (AI) is increasingly prevalent in our daily life. By feeding in huge amount of data, AI enables machine or computer programs to “learn”, analyze and perform complicated task independently.  Healthcare is one of the sectors where AI can be fully utilized on various platforms such as administration and medical diagnosis. Current applications of AI in healthcare include detecting cancerous tumors, neural abnormalities and cardiovascular abnormalities in patients as well as distinguishing high-risk leukemia patients. Now, researchers are looking into implementing AI in the surgical field, but at the same time many ethical questions and concerns were raised.

Positive ethical arguments for autonomous AI surgeons

Currently, there are already many cases of successful AI implementation in surgical procedure including the robotic eye surgery by PRECEYES and robot-assisted super-microsurgery by MICROSURE. From these positive outcomes, it is suggested that aid of AI improve surgeons’ performances, where technical difficulties (small, precise hand movements) and physical limits (performing difficult operations for long hours) could be overcome. These not only benefit the surgeons, patients could enjoy a faster recovery and shorter hospitalization – minimizing the crisis of bed shortage in hospitals. AI robotic surgeons were also proved to outperform human professional in both diagnosing disease and performing surgical procedures in a more accurate and precise manner – which is crucial when it concerns life and death. Utilitarianism argues that we should follow the act which will result in greatest net increase in happiness. So, although massive data is required for successful implementation of AI, this could be offset by the great increase in efficiency of surgical performance – which is definitely the desirable outcome for the majority stakeholders including patients, surgeons and hospitals.


Figure 1 AI robots can be more precise and attend to details compared to a human hand. Source: http://www.skzc.net/sort13/10168967.html

Various emergencies may occur during surgery and any possible outcomes highly depends on the critical immediate decision of the surgeon. However, a human’s moral compass is influenced by the seven emotions. From the virtue ethical perspective, it is difficult for a human to act rightfully virtuous consistently in their life compared to an AI robot, which can be programmed to “learn” the “ideal ethical principle” to effectively become more virtuous than humans themselves. It is especially difficult for a surgeon to control their emotions when facing some ‘exceptional’ patients. According to Surgeon Paul A. Ruggieri, he was once asked to operate someone who just killed his wife .Inthis case, he appeared to be devoid of any feelings, but he said ‘I wanted him to die’. Surgeons work in a VUCA (volatile, uncertain, complex and ambiguous) environment, so it is difficult to ensure that every surgeon can handle the conflict between personal emotions and the scalpel well. AI surgery robots however, without personal bias can treat patients as equals completely, which block out any outside interferences.

Negative ethical arguments for autonomous AI surgeons

Implementation of AI in surgical procedure may provide many benefits to the surgeons. However, how about the patient’s right to privacy who require to share their data? Data is the oxygen of AI. AI works as you must initially feed in massive data for it to “learn” and perform optimally. They require data to train algorithms so that the surgeries can be performed accurately, as any programming errors could harm the patient’s life.  Kant argues that we should respect the autonomy of others, treating them as ends in themselves and not as means to an end. The use of AI has neglected the Kant’s ethical theory, as it could violate the patient’s data privacy and protection. Are you sure that the data are stored securely? Who has the access to them? AI does not respect the autonomy of others when patient’s privacy is compromised as their personal data collected are used in means to an end to further the organization’s self-interest. According to a consumer survey performed by Accenture in 2018, 23% of patients indicated that they would not use AI tools since they are not willing to share their data.  AI technology has been a nightmare among patients who are concerned about their privacy and security due to the higher risk of cybercrime and loss of data.

Besides, cost of implementing AI in autonomous robotic surgery is very high. Although the price is currently unspecified, it is estimated that medically approving a partially autonomous medical robot with the Food and Drug Administration (FDA) will take around $94 million and 54 months to process. The overall cost to develop an autonomous medical robot will definitely be much higher, and this price will reflect on a patient’s surgical costs. Not many people will be able to enjoy this privilege, especially when it is reported that 83% of the world population in 2017 has low and middle-income range and may not be able to afford it. This would contradict the original intentions of the development of AI for medical surgeries in terms of utilitarianism, as only a small part of the world population will benefit from it and not all hospitals can purchase and use AI robots for medical surgeries due to limited funding.

From the virtue ethical point of view, it is unknown whether the intentions of the programmer of the AI is rightfully virtuous. Regardless of whether medical doctors are involved during the process, the risk of the AI being programmed to be biased towards monetary gains for hospitals will lead in negative implications, such as death of patients. Aside from putting medical surgeons out of jobs, the use of AI in autonomous robotic surgeries may also lead to skills decay among surgeons due to lack of practice, resulting in dangerous situations in cases where the AI robot failed and had to be taken over by a human surgeon.

Initial Decision

The use of AI in medical surgeries is deemed practical, especially from a technical perspective. As engineers who strive to apply scientific knowledge to meet social needs, we would recommend the development of AI use for medical surgeries. If moral values and ethics can be defined, it is possible that one day ethically aligned AI system can be developed to counterpoise existing ethical concerns.

53 thoughts on “Surgeons of the future – AI

  1. Very good argument for both sides. One greatest issue you stressed about is data privacy. However, most technologies nowadays have this issue – but are people actually THAT serious about their “data privacy”? Everyone uses Facebook, everyone takes photo and uploads them on Instagram, everyone uses the internet banking, which expose so many of their data privacy (unconsciously, for most of the people). Do people actually care about it? We are aware of it, and the consequences – we are aware of the possibility of data leaking and hacking, but we still use it. I personally think the concept is pretty much the same, it’s probably just the scale which is different. Also, it make need some time for people to actually accept this. The first time internet banking was introduced people were also worried for a lot of things – but look how much it grew. The same can happen to AI.

    I do very much agree on your point that we will never know the intention of people programming the AI robots. In this case, who should be the one responsible to it? The programmer? GM of hospital? Surgeons? Is it based on how much the hospital can earn? how many patients can they save? who will regulate them? these are the questions which really we have to think about and regulate it, before implementing this whole AI surgical thing

  2. Great argument that you raised on privacy data but a bit confused how does this correlates to Kant theory. You mentioned that data are required to feed into the AI and I presume that these data are such as blood type, blood pressure, etc, which can be easily be determined through blood test instead of personal information such as name, email and contact number. As long as these information are not provided into the AI (which I believe that is not required), how does this result in cybercrime?

    In the other hand, I doubt that implementing AI in surgery would result in higher price paid by patient. If so why would people want to build it at the first place as verified surgeons are known to be reliable. Marketing of this technology would really be an issue as patients will stick with highly skilled surgeons.

    1. Hi Seng, thank you for your comment. There are a lot of data available which if fallen to the wrong hands can cause unfavorable implication (such as dates of birth, insurance policy information, service dates, medical record numbers, and account balance). For the Kant theory, we are mostly arguing from the point which the data collected will be used in favorable for a third party’s interest – which does not respect the autonomy of patients – violating the Kant’s theory.

      Price wise, yes, in the long term it might be cheaper to implement AI surgery. But, how many hospitals will be financially capable to support the start up, research and the whole journey of implementing AI surgery? So this may potentially end up reflecting on patient’s bill. I am confused from your standpoint why won’t patient be paying the high price?

  3. In my opinion, using AI in the surgical field triumphs in both professional and ethical perspectives. Like the article mentioned, it has produced higher accuracy/success rate (which means probabilistic wise saving more lives in the world). Also, because robots are programmed with the ideal moral values, it wouldn’t be swayed by negative emotions and irrational decisions that would affect the outcome of a surgery. No matter how immoral the patient was in real life, everyone deserves an equal right to live, and his punishment shall be determined by court and not by any individuals.

    Speaking of the negative sides, it is certainly true that algorithms need to be trained but so do humans. There seems to be a double standard if we magnify errors by robots but not those by junior doctors. Medical errors are a huge problem but nevertheless we should hold both human and machine to the same set of standards. Moreover we know certainly in the long term, machine error rates would reduce significantly but we can’t say the same for human. In terms of privacy, medical records nowadays are stored digitally, that is irrelevant to the fact that whether robots or humans are the one performing the surgery. And we certainly can’t go back to the days where all data is stored physically. So then, it shouldn’t even be an argument in this case since data privacy issue has to be solved regardless.

    Yes, we might see a disparity in terms of people who could afford AI robots to perform their surgery. But if we think about the longer term, as we could see with all other technology in the world, price decreases over time, if machines are trained well enough to perform surgeries, then the programs certainly can be duplicated. In other words, it is more likely than not in the future, most people will be able to afford cheaper and better surgical procedure.

    1. I certainly agree with this. Especially the point which we need to train algorithms but human as well. It would definitely be easier to train algorithms so that all AI could perform well enough at a certain standard but that would be difficult to achieve if we were to train human.

      I think a lot of people misinterpret the “data privacy” issue you were trying to highlight here. Correct me if I am wrong but I guess what you are trying to bring out in this article is that it wouldnt be ethical for hospital to use the data of patients (their surgery cases, for example) for their own benefit (train more AI which eventually leads to greater income).

  4. Very interesting read. Have you considered interaction between patients and doctors? How do you build the fundamental “trust” between patients and doctors or surgeons who are operating them if they were robots? Can people handle their lives in the hand of a robot – where there will are possibilities of malfunction and errors. If AI robots dominates the surgical field, how do you convince people to this idea which is disruptive to norm?

    1. Hi Pam, thank you for the comment. As we are discussing about surgical we believed that interaction between patient and surgeon might not be significant compared to other healthcare perspective (ie nurses, physiologist or doctors). Personally, I don’t think I need to connect emotionally with my surgeon (I might need that with my doctor), but I would trust him if he have the right qualification and experience – which leads us back to our first point where we mentioned that robots are proven more efficient than human.

      It may not be easy to convince everyone but that is a normal process when it comes to introducing new technology. I believe with more successful cases and the right education about AI, people will eventually accept it

      1. Yes you are right about people will eventually accept new technology, but it may not be as easy for this case because what we are discussing here is relevant to life and death.

        1. Hi Pam. That is a good point. So, I think we could think from the point why would you trust a human surgeon more than a robot when they could both achieve the same thing?

  5. I certainly agree that the implementation of AI could cost a lot more, which would add to the patients’ financial burden. Hopefully as this technology gets more advanced, prices will drop to an affordable level.

    However, I have some doubts about the possibility of skills decay among surgeons. Even though AI is capable of increasing the productivity of a surgery by taking some job away, it can’t replace a professional doctor’s surgical experience and intellectual judgement.

  6. I do agree on the positives of having autonomous AI surgeons as I believe the pros outweigh the cons in the long run. For example, AI will always have higher precision with the micro movements, not susceptible to fatigue like humans do (which I believe increases the productivity and outcomes). On the topic of AI being more costly, this is only the beginning to the future. Yes, it may be expensive now to produce, but as it gets more common and more accessible due to the availability of this technology, I do not see the cost being a deal breaker. The traditional route of having extensive medical teams for surgeries might not be cheaper than having a well programmed AI.

  7. I like how you used the case of Surgeon Paul A. Ruggieri to support your argument. However in this specific case the surgeon in the end managed to control his emotion and operate his patient. Surgeons been through training and different experiences for them to act professionally, and could these be programmed in robot? If so, how?

    I do agree with the other comments that data privacy and cost shouldn’t be much of an issue, especially with the advance of technology these days.

    1. Hi, thanks for your comment. If we are talking about controlling emotions, robot would not even have that issue from the beginning. one big pros of using robots!

      The good thing about AI technology is that they could “learn” not from experience, but from data. These data could be past surgical cases which I would say equivalent to “experience”. I hope I don’t sound confusing

  8. An excellent article. I agree with your point on data privacy, which is quite a hot topic recently (Facebook data privacy case). However, I also think that it is a ‘necessary evil’ to share the data to allow the advancement of AI technology in robotic surgeons. As you said in the article, the machines need a lot of data to train algorithms before they can be allowed to perform surgery. Without it, it will be the same as a surgeon without knowledge. Therefore, it will be necessary for the corporations that develop the machine or the hospitals that operate it to get consent from the patients to use their data. The developer also needs to make sure the data acquired are not leaked or sold to other companies such as in the recent news about DeepMind Health that have access to NHS patients’ data is being transferred to Google, which is not a good sign.

    Another thing to discuss, will a human surgeon do better compared to an AI surgeon when an unexpected case occurs? In my opinion, AI surgeon nowadays still could not make an independent decision on what to do if an unexpected problem occurs such as a drop in blood pressure or haemorrhage during operation unless it was programmed with the data in the past. There are many cases where a surgeon had to solve an unexpected condition on the fly and saved lots of lives. What is your view on this?

    1. Hello, thank you for your comment and the interesting question raised. Personally, I think this is where the great of AI comes in. Imagine there is a big data hub where surgeons/hospital all over the world feed their past experience and data into it and this serve as a platform which the machines “learn”. This could potentially create a robot which has more “experience” compared to any other surgeon, and with that it could generate more accurate and possibly the most appropriate solution for any emergency situation (I am not AI pro, so I am not certain if this is how it is supposed to work, but yeah).

      So what is your opinion on this? Would a robot of such kind or a human surgeon be able to make a better decision at any emergency situation?

      1. If the AI surgeon have vast knowledge and experiences of past surgeries, I would say the AI surgeon would be better without a doubt. Data centre that collects all medical data is a good idea in developing robot surgeons but it is high probably not feasible in terms of data sharing consent and security in my opinion. Everything have pros and cons and AI surgeons are no exception.

        Having data accumulated in one big data centre will take a lot of effort and makes the organisation handling the data able to monopolise the service. Data breaches in said data centre could be devastating compared to data breaches of smaller data hub in a pool of data centres.

        Nevertheless, having all the data in one data centre can be convenient and easier to manage or access for further technological development.

  9. Such a well written article ! I do agree that the breach of data privacy is a real concern. However, given that lives of humans could be at stake, data privacy could arguably be a small price to pay ! I am also a firm believer that surgeons have the responsibility to keep improving themselves, this includes to be able to operate advance machinery in order to safe lives ! The traditional method of surgery has to change from time to time like it has always been. A really comprehensive article which addresses all the relevant issues of the field of A.I. Wowz !

  10. Fresh argument, well said !

    But if I have to choose, I’ll be on the side against the idea of AI surgeons. This is based on the point that the final cost for a surgery will surely expensive and only affordable by few communities which then against the idea of helping and reducing burden of patients. However, I strongly support the use of technology or robotic in diagnosing the patients as they are reliable and provide a fast and efficient results.

    1. Hi jiwon, thank you for your comment. I totally agree with you that the use of AI in medical surgery could burden the patients regarding the final cost for a surgery, as the initial implementation of AI surgery is very high due to the research that need to be done by the hospitals and its implementation is not easy. Hopefully, the cost for the implementation of AI surgery would be lower in the future so that more people would get the benefits from it and would be able to afford it.

  11. Could be potentially beneficial to doctors in performing difficult operations.

    However it should be a tool to help doctors, not replace them as loss of jobs would negatively impact economy.

  12. very clear and spot on arguments.

    Personally, I do support the use of AI in the medical field. Data privacy could be an issue, but I believe if it’s well regulated eg data protection act, it shouldn’t be much of a problem. For cost issue, I believe once the technology is well developed the price would stabilize and it wouldn’t take too long considering the growth of AI these years.

    Not only AI can improve efficiency like what you stated, I believe we could work towards challenging the impossibles and limits of human in the medical field with the help of AI.

  13. Hi, I love the contents presented and it is very well written. in my argument, I will be more biased towards the surgeon than AI. I doubted on the Al to perform surgery on the patient. in order to perform surgery, I think that hand skills and sense of touch are very important factors for a surgeon. how can we believe that the Al could perform the surgeries without having that sense of touch? because the human body from internal organ to outer skin is not something that can be systemic into the data system of the Al and it also varies for everyone. so, the input data of the Al might not accurate to certain people of similar surgery case. if there is some saying that it is not required to have a sense of touch for surgery, how can you convince people that the Al could work efficiently as a professional surgeon? moreover, it is also a waste for a student who studies for years to become a surgeon and it only takes a second to close their career pathway by just preparing a million of money to purchase Al. I am strongly against the functionality of Al as a surgeon. however, the implementation of al to assist the human limitations in other job scope is highly recommended.

  14. A very interesting subject. Implementation of A1 could be breakthrough in the medical field if proper controls are set to preserve confidentiality and security of private informations and protection against cyber crimes.
    Implementation should simplify the work of doctors and surgeons, not replace them.

  15. After reading the article, it clearly stated that AI has outperformed human in certain tasks. Article like this is amazing and makes us thinking what else AI can do. I can see how the easy tasks like making a coffee and classifying an object be easily replaced or outperformed by AI, what fascinates me the most is how the AI is able to outperform a surgeon in task that complicated like eyes surgery. But It takes a surgeon many years of experiences and practices to do well and make no mistake in an operation. An operation always seems to involve more than one surgeon. One might wonder if it will take more than an AI to perform a task like surgery, each specialising in doing what it is trained for. For instance, one AI will detect which is the affected area and another one is in charge of tumour removal. Or who knows, it could be all done by a single AI. Besides, How long will it take for an AI to train to the point of it doing at least as good as the surgeon? Few days or a few years? I’m guessing it all depends on the computer or server it’s trained on.

    If that’s the case, I’m wondering how much the companies are saving by replacing all the surgeons with a single AI or more. I believe hiring a surgeon can’t be cheap given the skills they are equipped with. In the end, if every surgeon is replaced by an AI, there won’t be anyone who is willing to pursue in the same field where he/she is outperformed by an AI. So when the AIs decide to turn their back on us. We will all be dead!!!!!! Because no one will have the knowledge to perform the surgery………

    1. Hello, thank you for your comment. Very interesting point you raised there! What is your standpoint on this issue then? Would you prefer the use of AI because it would be more efficient, or you would rather not have them to replace all surgeon in case they decide to turn their back on us? Could you also elaborate how they would “turn their back on us”? 🙂

      1. I think what he/she meant was it would be easy to manipulate robots than human. Imagine you have a batch of trained AI but someone with bad intention hacked or controlled them. This is a total possible scenario, that is why we have cyber security. This would not happen however if human surgeons were in action. You may be able to manipulate one human, but it would be difficult to manipulate a group of human at once. This would not be the case for AI, especially if all of them is under one whole system.. hence the risk. This may invite crimes opportunity, which the magnitude of it would be greater because it would be concerning live & death

  16. Ahhh. Another nice and interesting subject. AI seems to outperform human in many tasks. But what is the benchmark to say that this AI has outperform us in the task? For example, if I’m not mistaken there is yet an AI that could pass the Turing test designed to test the intelligence of an AI. It’s just a test to measure if the AI has reach to an intelligence level of a kid. If AI can’t even pass the kid test, who would believe that they have surpassed the surgeons in doing what they are good at? I wouldn’t let an AI that is not at least with kid intelligence to operate on me to be honest. What if it made mistake and killed me? Who is responsible for my death? The hospital, the programmer or the AI? Will the AI go to jail? Can I sue the AI? If the AI is programmed to perform such specific task, does the programmer require to have the knowledge to perform on the specific task as well? The point of AI outperforming surgeon in their job alone is not enough to persuade me to hand in my life to it. I think this does not only apply to me, in my opinion, if we want the people to believe in the “Robot”, there must at least have a strong evidence that is good enough for us to trust it in life or death situation. Else, sorry, I furthest I can accept is for it to determine if my tumour is maglinant or benign. Regarding surgery, I would still prefer an experienced surgeon than an AI that can’t pass Turing Test.

    1. Hello, thank you for your comment. Those are a few good points which we certainly have to take into consideration! I’ve not heard about the Turing test so I can’t comment much on that. However, if you follow the links in our article above there are a few interesting cases proving how AI surpass human ability in a safe and efficient manner in the surgical field. I’ve also came across this article about AI robot passing the medical license examination at https://futurism.com/first-time-robot-passed-medical-licensing-exam.

      I do agree a lot arguments can be made around the question who is going to be responsible for the mistakes made by robot, this could be a good reason to not implement the whole AI idea

  17. Hi. It’s so cool when we figure what an AI is capable of doing. Since the AI learns from the data, I believe the quality of data is one of the key factor that determines the performance of the AI. AI learnt from bad quality data might not be able to perform well. Hence it is important to ensure that the data is as good quality as possible. In data collection, we know that many of the data is not reliable. For example, some data is obtained through survey. Many of them are not answering the survey questions sincerely hence the data will consist of large amount of data that appears to be inaccurate. If the AI is trained with these data, it will definitely not work as expected. Moreover, if data were collected from measuring devices, a malfunction measuring devices will give an inaccurate measurement to leading bad data for the AI. So in my opinion, a measurement on how accurate the data is is a very important step. If there is no measurement, a doctor might alter the data to look bad hoping that the AI woundnt replace his job.

    1. I think Jackie raised a good point here which we might easily neglect. What if the AI trained do not work as well as predicted? Would that lead to more malfunctioning devices which would eventually cancel out the good points? Additionally, we will now be dealing with human lives, not like your smart lock or windows where the worse case scenario is you get locked out of the door… Not a good example, but you get the idea. If anything goes wrong here a human life would be taken. And in the end of the day, who can we blame?

  18. I definitely agree with the point where AI is not influenced by its own emotions. It treats everyone the same. It does not feel any pressure when performing surgery on VIP like the president or angered when it’s done on a rapist. AI is very consistent. If we have a copy of AI that make almost no mistake. We can duplicate it and make thousands of copies of it. We can do that to the top surgeon in the world. There will only be one top surgeon but there can be dozens of best AIs. With all these copies of AI that outperformed the surgeon, they save lives the surgeons couldn’t. Moreover, AI does not feel exhausted. They work 24 hours. Emergency operation happened during the midnight can be done by the AI. The hospital is always prepared for any unexpected emergency that will happen in any time. More surgery can be performed in day, we don’t have to be worried about having to wait for a long period of time after booking an appointment for a surgery. Hospital will never have the problem of lacking of surgeons or doctors.

    I understand that the cost to build the AI is very expensive. But think about this. All we need to do is to build one really good AI. And the rest of it will be copy and paste which gives you more than one good AI. I can imagine every hospital has their own version of AI because every hospital has its own kind of data. People will then opt for the hospital that has the best AI. I believe in the future it will be a competition between AIs and to find out whose AI is the best at solving their problems. Can’t wait to see it happening!

    1. Hi, thank you for your comment. You raised a few good points in supporting AI, also liked the idea of competition between AIs. As you said cost of building AI is very expensive, but the return of this may be great as well — taking your point of AI could more 24 hours, leading to serving more patients and reducing the waiting time. So if we consider in long term it might be a cost effective solution!

  19. An interesting article, and one that has generated a lot of comments already.

    Have a look at Care Ethics, which deals with relationships, since you are contemplating replacing the patient-surgeon interaction with a patient-AI interaction. The relationship here may be between the patient and the hospital.

    1. Thank you for your comment. We greatly appreciate your suggestion for other ethical theories to consider.

      With respect to Care Ethics, the main relationship that I can think of is the patient-surgeon relationship itself, especially since this is the most prominent relationship when considering a medical perspective and the fact that the patient’s life is at the surgeon’s hands. I agree that Care Ethics play a big part here as there exists a feeling of trust by the patient to the surgeon and a sense of responsibility by the surgeon to the patient, thus forming an interpersonal relationship based on care and trust between the two parties.

      However, I am a bit unclear on how Care Ethics can be implemented towards patient-hospital relationships in this context. Is it by evaluating if patient-hospital relationship can replace patient-surgeon relationships and provide the same level of care as the latter? Hope you could clarify this.

  20. In my opinion, if there is a development out there that is proven to be of greater efficacy and superior in terms of performance, it is only going to be a matter of time for it to be implemented. However, such leaps in developments at times places societal benefit over an individual. As mentioned above, not everything comes perfect and it is right to improvise and make changes (for example: data protection efforts, surgeons’ training etc) in order to minimize the negatives and maximize the positives. Ideal situations are not realistic. It’s about a slow and steady change, where people are allowed time to adjust and unforeseen loopholes can be rectified one by one. It’s part and parcel of life that every change, regardless of the best of intentions, is understandably to be met with an amount of resistance proportional to the change.
    On the bit where dangerous situations arises due to AI malfunctioning and human surgeons having to take over despite minimal exposure. It’s difficult to expect a 100% success rate even with robots as there are circumstances just not within our hands. Definitely by all means reduce it. But its about trading a 0.5 mortality risk by humans with a 0.001 (not real statistics) by the robots. Both causes of death equally tragic thus going for the method which gives a lower mortality rate may seem to be the lesser of 2 evils for me.

  21. Thought provoking article. I love how each argument is fleshed out.

    I agree with some positive arguments made, such as AI having high accuracy in surgical performance that outperform human professional, able to decrease technical difficulties and physical limitations of humans. I also like the argument of faster patient recovery and reduce in hospitalisation costs. I also like how the virtual ethical perspective is highlighted, where AI can be programmed to have ideal ethical principle compared to human.

    I also agree with the negative arguments made in this article, especially in terms of data privacy and security. The question of who has the access to patients’ data privacy and who will be benefits from it is a very important question in this aspect. I see some comments saying data privacy is a trivial issue, but to me, when it comes to ethics, data privacy and security is one of the biggest issue to discuss further when considering AI in the health industry. This is because while data privacy is deemed to be “normal to be shared”, how much is a person is in control of their data privacy? Especially in the medical perspective where all your details from A to Z such as your address, your bank account details, your medical insurance and other detailed background information is involved. This is the extent of information that hospitals take, and would someone risk all this information being available for everyone in the world to see and make use of it as they see fit? I think this may be one of the aspects that not many people realise the threat that may come about from the issue of data privacy.

    High cost implementation is also highlighted in this article, with approximately $94 million to develop it. This brings the issue of limited access to the AI surgery robot since it will be costly and only accessible to certain develop countries. I know some people might argue that with the progression of technology, this cost may decrease, but another question would lie in how fast will this cost decrease such that it can be beneficial to almost all people in the world regardless of their financial status.

    Use of AI can also reduce human interactions between patients and doctor. I believe human interactions is vital especially when trying to deliver the point or news across for example death of the patient during surgery.

    Aside from that, what will happened if there’s AI malfunction in the middle of the surgery? Who will be responsible? How will it impact the job employability in the future? Will medical surgeons require certain skills to control/ monitor AI?

    In conclusion, this seems too good to be true because the evidence for the success rate is still too few to be acknowledged and to be proven as beneficial. Instead of hastily implementing AI as surgeons, I think it will be more beneficial to use it in conducting research and data analysis.

    1. Hello, thank you for your comment. Happy that you add on to our article elaborating your opinion on each point. Someone in the comments earlier pointed out that data privacy such as bank account etc is in fact already stored digitally which makes it rather irrelevant now. Hence, cost may be the major drawback for implementing this, also with the few points you mentioned in the end. Can I ask if you have a clear standpoint if you support or not the implementation of AI in the surgical field, and your major concern behind it? 🙂

  22. I definitely agree on the implementation of AI as the surgeons of the future. Humans will make mistakes, make poor judgements, be easily distracted and have bad memories. Humans have limits and it is the AI that can help push them beyond that. The advancement of AI will save lives!

  23. Interesting article. great points from both side.

    I think a lot of people misinterpret the “data privacy” issue you were trying to highlight here. Correct me if I am wrong but I guess what you are trying to bring out in this article is that it wouldnt be ethical for hospital to use the data of patients (their surgery cases, for example) for their own benefit (train more AI which eventually leads to greater income). But again from your first point, if would benefit more people with the usage of this technology, and using their surgical record would be a trade off. Personally I don’t see a problem with this. If my surgical cases could be used as a case study to benefit and educate others and eventually save lives I would’t mind at all. But of course we could not expect everyone to have to same mind – hence the ethical debate

    1. Hello, thank you for your comment. Thank you for clarifying that is one of the point we would like to stress about but perhaps we are missing a few words on our article. So from your comment, could I interpret it as you would support the implementation of AI, despite the needs of using your data, ie surgical record as a case study?

      Personally I would not mind as well. I believe most people would not but as you said we could not expect everyone to have the same mind.

      1. I think it is not only the matter of if one would mind his/her surgical case being used. Think about it, if many people are not willing to opt in, we will be missing a lot of data for the AI to study. Will the AI still be as effective then? Will missing data lead to any sort of bias? This will counter the whole point of having AI robots to aid surgery at the beginning

  24. The mortality rate of mankind has decreased significantly throughout the past century. Since 1900, the global average life expectancy has more than doubled and is now approaching 70 years. One of the major contributors to the increase in life expectancy is the advancement in medical technology.
    As indicated by the article, according to Utilitarianism, we would ideally want the implementation of AI surgeons to be on the mass public. However, a success rate close to 100% is unlikely to be achieved by AI surgeons, which could lead to many misfortunate surgery failures.
    What happens if an unexpected event, which has not been introduced into the training data of the AI, occur during the medical operation? Will that cause the AI surgeons to malfunction and go ‘berserk’, ultimately killing the patient? Is there a backup plan that can guarantee the survival of the patient without any human intervention? I can envisage the use of deep learning in computer games like Go, Starcraft 2 and Dota 2, given that everything occurs within the game itself. Is the use of AI surgeons feasible given that there are so many parameters in real life surgery to be ascertained?
    I appreciate biomedical technologies such as X-rays, MRI and CT scan, which assist doctors around the globe on diagnosing patients, but currently I am sceptical with the use of AI surgeons.

    1. Hello, thank you for your comment. I do agree that it is a difficult process to gather all data possible so that AI could train on them to counter all possible situation. Something which AI may never top against human. Given there are too many parameters to reconsider, we may need a long time for a fully autonomous AI to be implemented in the surgical field. That was a really good point raised! thank you.

  25. Hi. Interesting article you have there.

    Personally, I would not be a fan of AI robot as my surgeon. Yes, human may be fatigue or experience emotions. But surgeon is a profession, they were trained so that they would minimize the effect of these limits when performing their job. AI robots on the other hand, could be exposed to cyber hack, malfunctioning, machine errors, the list goes on. The less control human have, the greater the effect of these drawbacks. Unlike human where they could be held liability and responsibility for errors they made, we can’t to that to robot. Will this be a loophole? Something to think about.

  26. Ai is bound to become the choice of more people because of its advantages compared to human doctors.
    AI must carry all the cases that can be verified, including various special conditions, and updated by the developers in real time. AIs around the world can receive up-to-date information at the same time, and AI takes much less time to absorb knowledge than doctors.
    Any surgery is risky, and the measures taken by AI based on knowledge reserves may not be worse than every surgeon.
    On the contrary, AI is not bound by emotions, and it is absolutely cool to deal with the situation scientifically. In the virtues mentioned in the article, AI is absolutely fair. (can be added to the relevant data processed by the doctor’s emergency)

    1. Agree with you. What’s more, the experience of surgeons is a two-edged sword. Beacuse experience is limited, the cases one surgeon know is limited, then it may cause wrong judge. But the data the robot get is much more than one surgeon, it;s more reliable.
      Therefore, it’s good to develop AI surgeon to clinical surgery.

  27. Data privacy is important, but the leakage is normal these days, like the Facebook information leak that shocked the world, the risk of data leakage is always there. There is no absolute security system in the world, but this is a problem that can be avoided as much as possible. Developers will take the privacy of patient privacy more seriously, such as developing a more secure and private storage method. In addition, if extreme conditions occur during surgery, rapid patient information may be more helpful for surgical judgment.

    1. I think data privacy is really important, espacially the information of our own health. Privacy is a necessary condition for an intrinsic value. The loss of privacy would therefore be a threat to our most fundamental values.

  28. I support the use of AI roberts as surgeon. As for the problem of cost, it will not exist with the development. The application of AI in surgery will be gradually developed, improved and popularized like a computer, and become a product that everyone can consume.Every development of technique has such period to realize the wide application.

  29. If the computer center was hacked, a normal surgery could become a murder. What if the program has fatal flaws or bugs, no wonder patients could confront more risks this way.

    1. Definitely agree with it. If this do happen, then this technology can be mean to plan a murder. The risk of the surgery may not be reduced, but rise. What if apply robot just in some certain situation, like battlefield.

      1. Agree. Love the idea of only applying robot under certain circumstances, like war or battlefield – where usually there is way lot demand AND very limited supply of surgeons. Risk of being hacked should be taken seriously. As far as I am aware, the more autonomy the robot have, the more security risk it poses. Where in the case of human surgeon, very unlikely he or she can be controlled or manipulated by others.

  30. Use of AI will invite many risk.. hacking, lost of telecommunication, system failure, robots have their own limit too. If you google da vinci robot there were many articles on how they work well, but also on the technical errors they had. I don’t think we should depend too much on AI, unless it is for overcoming something human could not achieve, for example too tiny hand movement

Leave a Reply