They help the disabled, entertain patients in hospitals, and support therapies for autistic children. But from robots in healthcare, we expect much more than just manual help. We want human touch. Humanoids augmented with social abilities should react to our emotions and create authentic relationships so we can trust them. What care do we want in the future? And how can we involve robots to make healthcare smarter?
A patient full of emotions
A small, white robot with big eyes enters the patient’s room. It asks questions about the mood, takes a thorough medical history, carries out necessary laboratory tests. Artificial Intelligence system analyses all gathered information. A diagnosis is ready in a few seconds. The patient receives a prescription, an exhaustive list of recommendations, and a bill. Theoretically, an excellent patient experience, perfect service where evidence (data) based medicine is paramount. But even the best robotic doctor is not able to satisfy the patient’s emotional needs, which are equally important as treating disease. The machine does not show emotions, does not have empathy, and it does not respect the patient’s perspective. It just follows the procedure.
Although robots perform well on a production line, interaction, communication, support, and empathy are more significant when it comes to medical care. It is more essential to meet the social needs of the patient that are one of the most significant ones in Maslov’s hierarchy. Furthermore, they are the most difficult to satisfy, as they are intangible and elusive, and they vary among people. We realize social interactions differently, just like we alleviate hunger through food adjusted to our taste and culture. Of course, as long as we can afford this luxury. Everything depends on the living environment, upbringing, personality, and priorities. While experiencing disease, all emotions present in a human being are brought to a boil. In this kind of situation, we look for understanding and help, and not only a prescription or best treatment.
Social robots are getting better and better at recognizing and reacting to our emotions. Voice assistant Alexa is already capable of intoning utterances, expressing feelings, and communicating with a human being in a better way. So far, these are only two emotions, i.e., excitement and disappointment. In a few years, Artificial Intelligence will learn a more natural language and voice modulation, so that we will not be able to tell the difference between a chat with a friend and a chat with algorithms. Cameras built into robots, observe with the highest precision, and analyze expressed feelings so the robots can adjust their behavior and reactions to the situation.
Robots can gain our trust only by behaving like a human being. We want them to be able to talk with us and to treat us individually as well as arouse various emotions. The number of innovations which can do this is increasing. Paro, a cute baby seal robot, is a favorite among patients suffering from Alzheimer’s disease and dementia. There is an authentic relationship between Paro and patients, very close to the one between a human being and a real animal. Moxi, which was designed to help nurses, became a celebrity for patients who treat it as something more than a soulless machine. Pepper — the world’s first social humanoid robot able to recognize faces and basic human emotions — already works at the hospitals’ receptions and help in the therapy of autistic children.
The social skills of robots are the subject of many elaborations. In the biomedical literature library PubMed, the term “social robots” gives 1,400 results. A majority of abstracts examine the potential of their usage in the care of people with special needs, such as patients with neurodegenerative diseases, or children suffering from various diseases. One can also find there critical texts, which draw attention to the dehumanization of social contacts in healthcare, which is the beginning of the epoch of post-humanism.
Binary healthcare with compassion
Technology is never neutral. Robots with social competences will also have far-reaching consequences, affecting society, social relationships, and the definition of a human being. Intelligent machines that can recognize our feelings more precisely than another person seem terrifying but also fascinating. The Japanese invention called Gatebox, which is a virtual friend in the form of a hologram, may serve as a good example. This personalized avatar remembers things about its owner better than the best partner. It is curious about how we spent our day, it asks us how we are, and it is happy when we come back from work. Some people consider it to be a sad invention, which demonstrates the dissolution of social bonds. For some people, it as a great innovation, which helps lonely people who are not able to establish real connections.
However, today is not the right time to ask questions about what we expect from health care. When the question of what is crucial for us as people is asked, I always ask what we mean by ‘we’? Do we have a moral right to judge a lonely, bedridden person, whose only friend and assistant is a robot because the family does not have time for a visit? Is a companion of an ill child in the hospital in the form of the Huggable robot created by Massachusetts Institute of Technology MIT a fraud or support in difficult moments? Everyone values different things. It all depends on many factors and perspectives. Therefore, it is not fair to generalize questions about robotics in healthcare. It is wiser to consider what we want to avoid.
One thing we know for sure is that robots used in healthcare should not cause harm. Similarly to doctors, following the ethical principle of Hippocrates primum non nocere, robots should not collect information for purposes other than the ones concerning care over patients. They cannot act as a spy standing next to the patient’s bed. Algorithms controlling these types of solutions should be created transparently and ethically. Robots should remain emotionally neutral so as not to manipulate feelings or raise any hope that they can do more than in reality. However, there is a fine line. Does a machine that holds the patient’s hand or gives hugs already violate ethical standards?
At the moment, robots should not make independent and un-programmed decisions. The real world is too complicated for the perception of an average robot, and even the best algorithm of Artificial Intelligence is not able to analyze a given situation as precisely as the human mind, paying attention to the smallest details and nuances. Robots should also be neutral in terms of their attitude towards patients (lack of prejudices and partiality) and collect only such data that patients want to share.
Synthetic empathy or a virtual lie?
Bart de Witte, the Founder of Digital Health Academy, predicts that empathy generated by the Artificial Intelligence systems will become one of the leading trends in a few years. People establish relationships based on trust with machines when they understand our emotions and can react adequately. Today’s Artificial Intelligence is very limited. At present, many services use other psychological mechanisms in order to increase the use of a given technology. Social media, which is addictive because based on the simple principles of the psychology of behaviors and the stimulation of a reward center in the brain, are a good example. Therefore, each “like” generates a feeling of happiness, acceptance, and strengthens self-confidence. In a few years, Artificial Intelligence will enter virtual reality by imitating the behavior of a human being. Robots that act as doctors during telemedical consultations can be equipped with synthetic empathy. They will be able to respond to our needs correctly and will never criticize our behavior. Will it then turn out that a conversation with another human being, during which we often face criticism or another point of view, is too toxic? Think about how many patients change their doctors because “they feel misunderstood.”
“Technology can always be a threat or a chance for humanity. It is us — people — who define the purpose of technology. Currently, many of the business models digital companies use are based on digital services that compete for users’ limited attention. The modern economy increasingly revolves around the human attention span and how digital services capture that attention,” says Bart de Witte, the Founder of the HIPPO AI Foundation and Digital Health Academy.
“The addiction to social validation and bursts of “likes,” for example, is destroying our real-life attention spans. Our brains are drawn to outrage and angry tweets, replacing democratic debate. During the last few years, access to technology’s godlike powers has increased dramatically, while the ancient, Paleolithic impulses of our brains have remained the same. What happens when we add the power of love in the box of godlike technologies and give it to 25-year-old nerds, backed up by venture capital? Algorithms are already able to read our emotions much better than we humans do, but as they do not possess consciousness, they won’t have the ability to feel. Empathic behavior will be based on synthetic empathy, which will become the most powerful tool developers have ever had access to. Before we all fall in love and enslave ourselves, we need a more open debate on how far we need to regulate this,” Bart de Witte points out.
At this point, the question of whether we will shape technology or whether technology will shape us arises again. Inventions like synthetic empathy, robots with emotions or virtual friends are not as controversial as human cloning or editing genes about which most countries share similar ethical recommendations, introducing legal barriers. The influence of digital innovations is far subtler, long-term, and difficult to predict. Perhaps it is equally dangerous.
Experiment on patients
It is not possible to clearly separate roles which should be fulfilled in healthcare by robots and by human beings. Medical care provided, for example, by a nurse, is a mixture of technical and social tasks. It is also challenging to think about target models, as technology is changing dynamically, and tomorrow today’s ideas may be outdated.
The market does not stand still either. It experiments with the growing number of new models of robots implemented in pilot projects. Many of them are created by innovative startups, the ambition of which is to develop revolutionary technology and achieve market success, not to investigate issues like “values” or “ethics.” Most of these projects fail, but projects that pass the market test join other innovations and create the force changing current healthcare. A good example is the Moxi mentioned above, created by Diligent Robotics. The Zora Robot, which costs about $10,000, is one of the most popular humanoids. It is used in care centers, where it entertains patients, talks with them, dances, exercises, and makes patients laugh.
The evolution of machines-carers has significantly sped up along with the development of Artificial Intelligence and the so-called neural networks. Robots that accompany patients, and people who require special care, may learn their needs, habits and gradually adjust their skills. For example, the project GrowMeUp implemented between 2015 and 2018, co-funded by the EU Horizon 2020, was one of the most extensive plans in this area. Its objective was to develop a cheap robotic system that would be able to learn the needs and habits of the elderly and grow up with them. In this way, robots would be able to compensate the deterioration of the senior’s cognitive abilities and support them in everyday life so that they are active, independent, and socially engaged for a more extended period. Robots would transfer the collected data and knowledge to a data cloud so that they could learn from each other and create a combined virtual care network. Ambitious plans ended up with no success because we still lack the right technology.
Robotics in healthcare is like walking on thin ice. The lack of regulations leads to uncertainty. Moreover, such projects are rarely funded in public healthcare systems. Therefore, for manufacturers, it is challenging to create a stable business model, and they focus on the private health sector. This poses the threat that only more affluent people will benefit from robots. Interestingly, many experts believe that the situation may be completely different. If the mass production of robots results in the fact that they are cheaper labor force than people, more indigent patients will be served by machines, and only wealthier patients will be served by a real human being.
In the middle of extreme scenarios
Robots are a new labor force that is meant to fill the increasing gap related to the lack of doctors and nurses. It will be relatively easy to teach machines how to take care of patients, change the sheets, bring meals to bed, measure body temperature, and control the general health condition. Robots, whose competences will be reduced only to the foregoing skills, will remain only mechanical assistants.
If they perform tasks carried out by human beings, we want them to look and behave like human beings. Therefore, such robots as Moxi or Pepper have human faces, big eyes, and a broad smile. They do not frighten patients but make them feel more comfortable.
However, it is difficult to imagine that healthcare based on human beings could be exchanged for a model based on robots, in which machines take over all of the competencies of doctors and nurses, taking care of patients on their own. At least not at the moment, due to the lack of sufficient technical capabilities. When a few years ago, I read about the Paro robot, which helps taking care of patients suffering from Alzheimer’s disease, I was delighted by the idea. I was extremely disappointed when I held the artificial baby seal in my hands. Mechanical, clumsy movements, monotonous murmurs, and robotic eye movements turned out to be a very crude form of what real animals can do.
“We are holding a tool in our hands, which can be an opportunity but also a threat to healthcare. In order for technology to become the Trojan Horse introduced by companies providing new technologies for healthcare, we must get involved in its development, participate in research projects, discuss their application with patients, politicians, and all stakeholders” — global health expert, professor Ilona Kickbusch, rightly claims. It will take years before robots learn such basics as smooth movement around hospital rooms or starting a relatively neutral conversation with patients. Meanwhile, we have the time to focus on the development of the model of intelligent and comprehensive healthcare (“smart care” and “deep care”), where robots support people, assist them in mechanical and physical activities and leave the sphere of communication, empathy, and emotions to human carers.
Nevertheless, we know that human curiosity and willingness to experiment with Artificial Intelligence and technologies will still accelerate newer and newer models of robots imitating human behavior. And there will be nothing wrong with this, as long as they help patients instead of harming them.
Originally published at aboutDigitalHealth.com on the 25th of February 2020.