"So you really think robots can replace me?" laughed the family doctor. "How will they understand when your child is in pain? How did they know he wasn't just faking it because he didn't want to go to school? There are things only humans can do!”
Those of you who have tried to explain to doctors the idea that artificial intelligence could take over a large part of their work, have probably encountered similar reactions. Unfortunately, these are instinctive reactions that are mainly based on a partial understanding of the capabilities of artificial intelligence today - and completely ignore all the other directions of development that characterize the future.
Let's go through some of these directions now, because I believe that an overall view of technology today inevitably leads to the conclusion that AI engines will also be able to handle a large part of human interactions and attribute proper meaning to them.
First direction: the development of emotional computing
The more advanced the artificial intelligences are, the more they can decipher and understand emotions - sometimes based on a textual description only, and sometimes by viewing an image or video. An algorithm developed at the Massachusetts Institute of Technology, for example, looked at the faces of people who were in pain, and was able to rate the pain they felt[1]. Research from the University of California and Toronto also revealed that computer systems are even better at detecting pain than human doctors, and screen out 85 percent of imposters - while trained humans only managed to spot imposters 55 percent of the time.[2].
This field, by the way, is called Affective Computing, and it is expected to jump in market share from 9.3 billion dollars in 2015 to 42.5 billion dollars in 2020[3]. We will see emotional computing everywhere: it will be used by artificial intelligence so that they can function better with elderly people in nursing homes, children in class, drunks, gamers and much more. Oh, and in front of patients too.
So how did computers know that 'sick' children were faking? Much better than human doctors.
Second direction: information from different sources
I explain over and over again that in the not too distant future, we are all going to be monitored at all times. The sensors will be all around us, in our homes and on our skin and clothes. This will be the Internet of Things. There are predictions that by the year 2020 we will reach the number of 200 billion objects that will be connected to the Internet of Things. Even the most pessimistic forecasts assume that by the year 2020 each person is expected to own an average of 6.3 connected objects[4].
Human doctors are unable to deal with the information received from all these sensors, but artificial intelligences will easily be able to deal with it and cross-reference it to reach a final conclusion regarding the mental and emotional state of the patient. They will be able to know if the patient sleeps well at night, how he talks in his daily life, whether he enjoys going to school or work (according to his emotional state as measured by heart rate, sweating, breathing rate and other parameters) and so on.
So that once again - artificial intelligences will be able to decipher emotional and mental states better than human doctors.
Third direction: ability to imitate human reactions
The most intriguing conclusion from these two directions is that in order for a person to receive medical treatment and diagnosis in the future, he will not necessarily have to go to the hospital or the neighborhood clinic. It will be able to allow the computer to have access to all the information received from the sensors around it, and to answer computerized questions to get a preliminary diagnosis.
But how will communication be conducted with these computerized doctors?
Surprisingly, it may be very similar to a video call with human doctors as it exists today. The reason is that developments in the field of artificial intelligence today give algorithms the ability to create avatars - digital characters - whose reactions are more and more similar to those of humans. Facebook, for example, managed to create a bot that watched hours of Skype conversations between people, and then successfully controlled an avatar that responded to the facial movements of its interlocutor with appropriate facial movements of its own: when the person laughed, the avatar also widened the space between its lips, or tilted its head in simulated amusement . Bystanders judged the computerized and human avatars to be equally natural and realistic.[5].
So it is definitely possible that the computerized doctors will also be able to treat us - taking into account many of our emotional needs - via the Internet.
What will be left for the human doctors?
Well, what will be left for human doctors, given a future world where artificial intelligence is able to reach more accurate diagnoses and medical recommendations than the average doctor?
My first recommendation to doctors has always been to understand how to work with computers and get the most out of their recommendations. My second recommendation has always been for doctors to focus on developing social skills, so that they can provide what computers have a harder time giving: empathy.
But what happens when emotional computers will also be able to provide the patient with empathy and emotional support?
In this situation, there is only one great advantage left - perhaps the only one - for the human doctor: his humanity. Human doctors will always be more expensive than computer doctors. They will always need more work than the patient - to come to the clinic, sit in the waiting room and stand reverently in front of the human doctor. They can deeply impress the patients with their white coats and their deep and authoritative voice, with their firm handshake and the deep look in their eyes. All these are characteristics of treatment in the physical world, which it is hard to believe that robots will replace in the coming decades.
Or in other words - the last work of the human doctors will be in a theater play, in which they still pretend to have some importance in their opinion, and the patients participate as an audience of their own accord. The work of the doctors will be to 'deceive' the psychology of the patients. rely on their basic human cognitive biases.
And such work, apparently, will remain open as long as humans continue to be human.
As always, if you want to read about the estimated future of work, and the different possible models in which we may work in the future, you are invited to read more at "The guide for the future" and "who rule the future” – both in bookstores.
Sources
- [1] https://arxiv.org/pdf/1708.04670.pdf
- [2] http://ucsdnews.ucsd.edu/feature/computers_spot_false_faces_better_than_people
- [3] http://www.marketsandmarkets.com/PressReleases/affective-computing.asp
- [4] https://madaduhcom.wordpress.com/2016/12/16/%D7%A2%D7%AA%D7%99%D7%93-%D7%91%D7%AA%D7%99-%D7%94%D7%97%D7%95%D7%9C%D7%99%D7%9D-%D7%95%D7%A2%D7%AA%D7%99%D7%93-%D7%94%D7%91%D7%A8%D7%99%D7%90%D7%95%D7%AA-%D7%A9%D7%9C-%D7%9B%D7%95%D7%9C%D7%A0%D7%95/
- [5] https://www.newscientist.com/article/2146294-facebook-ai-learns-human-reactions-after-watching-hours-of-skype/
More of the topic in Hayadan:
5 תגובות
When the computer is controlled by humans, the machine can harm people, even if a hacker is behind the barricade. There is and will not be faithful as man to man. Then the doctors will consult with Mrs. Bina. that's it.
Supposedly an interesting article by "Roi Cezana".
He writes: "Unfortunately, these are instinctive reactions that are mainly based on a partial understanding of the capabilities of artificial intelligence today - and completely ignore all the other directions of development that characterize the future."
And I ask - Roy, could it be that in writing the article you yourself ignored a number of data that are not currently in your mind?
I agree that if you look superficially, the artificial intelligence will indeed be able to recite spells and medicine in a fast and purposeful way. Effective and maybe in a few decades also emotionally.
At the same time, a number of variables that were not mentioned in the article should be taken into account.
1. The depression!!! Depression threatens to become the number one disease in the world in a few years, according to the World Health Organization.
2. New studies have found that our happiness does not lie in money, health, nor status, but rather in a good relationship in raising children and relationships with friends at work and outside.
3. The difference between a computer and a machine, you can call it artificial intelligence, is that the machine is not able to create a new development, to see ahead. It is a machine that implements the intelligence found in it, which is also the product of an inventor.
Where does all this take us?
that if we are talking about typical diseases then I agree that artificial intelligence may provide an adequate solution.
If we are talking about what will really torment humanity in the coming century, what a disadvantage it will be that it can be associated with the cause, the root, the cause of all ailments. So no computers, no robots, and no artificial intelligence will help us.
We will need human intelligence!
We will weave a filling that will go exactly into those shortcomings that are still unconsciously with us. We will yearn for a new consciousness that will flow into the veins of our consciousness and provide us with a psychological compensation for all that we lack, but we do not yet know what it is.
All of this will inevitably lead to the discovery of a new method, a new paradigm that will be considered the method of the creation of humanity.
A lot of doctors don't have emotion anyway.... Already today you can build a software that will put all of your feelings into it and give you a diagnosis
You just need to enter all the data into a huge database and let the computer give you options with percentage chances
In some cases, a person who needs medical services is waiting in a section, especially in the periphery
It can last more than six months and even more in ping-pong between the family doctor and the specialists who, in some cases, except for tapping one finger on the keyboard, are barely able to see who is even sitting in front of them, so in advance there is almost no family doctor or specialist available (they haven't opened the diary yet - familiar to you), and if They are already available and in quite a few cases an AI system will be far superior to them
Even in the availability that it can already be the mobile phone from the capacity for diagnosis including recommendations
(For a blood test and injections you only need a nurse and maybe in the future it will also be possible to make it automatic)
Personal databases that will be accessible to these systems in addition to a comparison with global databases with the most up-to-date information available, including information cross-checking capabilities and speed on humans, where human touch is needed, clicks, etc.. where you will be referred to a doctor,
to be more of a control systems supervisor,
The reason that doctors are in the crosshairs of these systems is one of the huge cost and time of training a doctor
A constant lack of doctors, people suffer and some of them fall behind waiting for the ping pong until they get the long-awaited diagnosis.
The second reason we all suffer with it is we or a lost family member and when the people who develop the AI systems
They are looking for where to apply the developed systems. The first place they want to apply is medicine to reduce human suffering. Although medicine has undergone a huge change in the quality of life, there is still a lot to improve and one of the promising ways is the AI systems.
Interesting article, but in my opinion the final conclusion is wrong. In my opinion, the main work of doctors in the future will focus much more on the research and development and refinement of those bots. Being "actors" can also be a relative advantage, but woe betide us if this is the main skill they will have to impart in medical school.