Not in a long time, even if it might be helpful to have medical automated asistatants

Not in a long time, even if it might be helpful to have medical automated asistatants

Dear Enrique Garcia Moreno Esteva

so, do you think it should not have responsibility, but that would perform support activities, such as nurses, or litter-drivers, food carriers, and so on .... without formulating diagnosing interpreting laboratory tests, or surgery in autonomy

 Dear Tobias Mettler

You think in the "American way", but in Italy rich would want robots ... to be more advanced and ....cool, and the poor would have not too motivated physicians ... while the majority of experienced physicians would work in private practice......anyway, you formulated a valid observation, Thanks

A

From what I have read on topic so far, the robots are not yet ready to fully autonomously function in complex social environments with a lot of possible emergencies and exceptions. There are just too many things that could go wrong and are hard to foresee. Therefore,  using autonomous robots as nurses and doctors is not really feasible yet.

Autonomous and semi-autonomous surgical equipment might be another story, but as you mentioned, there are ethical and responsibility issues. So far the community is still trying to figure out what to do with these issues, so I would not expect a wide acceptance of such devices before there is a common agreement about these issues.

In my humble opinion the ethical and responsibility issues are no different than those involved in having autonomous machines operating in any environment alongside humans. Take the autonomous car and the unavoidable death thought experiment - What does it do when faced with a situation in which there is no outcome in which a human does not die, i.e. some one walks out in front of it? Does it continue and possibly kill that person, swerve left and possibly kill another pedestrian, swerve right and possibly kill a cyclist, or brake so violently that control of the vehicle is lost and possibly results in the death of its passenger? How does it learn the ethical principals that underpin the least worst choice? Can programming and learning (in simulated environments) replace true experience i.e. learning by making mistakes? There are very few Machine learning strategies that allow the autonomous machine to come to its own conclusion on what the "right" outcome is. Most are "trained" i.e. guided learnings that provides outcomes near to what the human controller desires.

The other more pertinent risk from a development point of view is in our litigious world, the law is trying in vain to keep up - will victims will sue the company/programmer or the machine?

Dear Gabriel,

Accurate observations. You are absolutely right, but given that malfunction in electronic driving assistance systems in contemporary cars (ABS, Brake safety, adaptive cruise control, etc) can create problems you mentioned, whereas healthcare domain is still unexplored on the issue. Moreover, it affects a greater number of people, due to health domain relevance. Finally, driving or using other electronic equipment, it is a choice (almost unavoidable, actually), conversely helthcare assitance is a right.

So I think that the two areas are a little different according to autonomous robotics assistance...and in any case I think that manufactures have to take on the responsibility... even movies teach that in case of uncertainty, security protocols must exist

I think we are talking about two different things here.

My example was not discrete component based, there are also many of those in medicine, i.e. heart rate and blood pressure monitors that malfunction at crucial times. It was based on the ethical considerations of an autonomous learning system. The example was given in the context of the autonomous learning system that is (or proposed to be) the google car.

Autonomous learning systems are self* systems that require no human intervention. The ethical considerations only come to the fore when something goes wrong. In medicine, which as a discipline is a "best effort" system, i.e. there are absolutely no guaranteed outcomes, every procedure carries a risk (and you as the patient sign the form accepting that risk), there is always a potential for something to go wrong.

The ethical question then becomes twofold - when the autonomous self learning machine (with adaptive dynamic code) is making the decision, who, if anyone, is responsible, the manufacturer, the programmer, the trainer when something goes wrong? Remember in a 99.999% reliable systems with 30million lines of code there are 300 different opportunities for something to go wrong and that changes with self coding machine learning systems. In a wider context the more important question becomes - if the benefit of autonomous is proven to have better outcomes than real surgeons is it unethical to withhold it?

Dear Gabriel

I agree with you, probably we are talking about two different things here, I was referring to the different "sensitivity" required for the healthcare issue, which as mentioned is a right ... if I do not want to use an "autonomous" vehicle, I can't use, but if my surgeon or caregiver was a robot, I have a right to object, if I were skeptical.

Nice exchange!

Sharing link;

http://theconversation.com/robots-in-health-care-could-lead-to-a-doctorless-hospital-54316

Regards!

Agree with Tobias Mettler

regards