Two American scientists, Michael Bannissy and Chris Edgar, have rolled up their heads to develop the first machines that empathize with people or other intelligences, human or artificial. But someone will have to teach the AI to be empathetic. This could be one of the jobs of the future.
It is already difficult to be it, or to achieve it, for the human being, as for above asking the machines, or artificial intelligence (AI), to be it. Empathy is, and I don’t look at any dictionary, that technique or connection that some people have with others, without knowing many times why it exists. Some call it ‘chemistry’, not to mention empathy. And now, neither short nor lazy, Michael Bannissy, head professor in the department of Psychology at Goldsmiths, and Chris Edgar, postdoctoral researcher at Banissy Lab, are thinking of talking about empathy in artificial intelligence.
What exactly are they referring to? Artificial intelligence (AI) is increasingly present in our daily lives, whether through the voice assistant on the phone or the complex algorithms used to win the battle against diseases. How we design interactions with AI systems and the results they provide must be carefully considered. Furthermore, in the future, the responsibility for designing artificial empathy could fall to a new job / position that does not yet exist: the “empathologist”.
Michael Bannissy and Chris Edgar have investigated how people perceive, interpret and interact with others, offering a unique and academic perspective that goes beyond the field of pure technology. Empathy, they say, is a fundamental part of our lives, contributing to collaboration and cooperation, and transcends cultures, social behaviors, and well-being. According to Banissy, “human beings are social beings. If you really want to think about developing technology like AI, I think it is crucial that empathy is involved in some way. “
An artificial intelligence system with the capacity for empathy could provide more natural interactions, taking into account our mood or feelings. “Depending on the type of information you want to convey, how you want individuals to react, or how you predict that they will react to it, it will be necessary to understand the current state of these individuals,” says Edgar. “In a way, their emotional state will give you a clue as to how they will respond.”
Voice assistants are one such example. The way they deliver information could be modulated to respond appropriately based on an individual’s current emotional state, resulting in a more natural experience. This, however, is not an easy task. While the machine may learn to differentiate between smiles and frowns, the reality is that we can cry with joy or smile with frustration, depending on our mood and our individual personality.
“Currently, machine learning algorithms are quite good at recognizing some physiological or behavioral signals that are associated with specific emotional states,” says Edgar. “However, if you are looking for the most sophisticated theories of emotions, and how they are produced, there is a great variety of them, and there is no correct one above the rest. They are all complicated, and it is not as simple as recognizing someone’s tone of voice, posture, or facial expression. There is much more than that”.
Someone with an academic background in empathy could use their experience to improve understanding of different emotional nuances by AI systems. At the same time, it would work on improving algorithms to better understand the emotional signals of individual users. Ultimately, it could help systems go beyond just measuring surface expressiveness.
Diversity is considered one of the pillars of teams to achieve success, since different points of view and experiences add up to achieve better final results. As a reflection of this, an artificial intelligence system is only as good as the model on which it is built – if it is fed with a reduced set of data, its functionality will be more limited. Therefore, the role of an empathologist could be vital. The different insights and insights they could provide will help ensure that systems are designed with a vision that transcends the purely technological.
Banissy says that “if you think about it more broadly, I think this is really someone who can bring the insights from a wide range of complex sources, to help AI development teams in the most efficient way possible. There are numerous different models and perspectives, and it is very easy to get confused quickly. It would be helpful to have someone with experience, who can help you understand the different components of what is most relevant. It’s nice to work with other people from different perspectives, because they help you find different approaches to tackling problems. That is the key to a highly collaborative and innovative team ”.
Designing a system with the help of a more diverse team and feeding it with a wide range of data will lead to algorithms that will be able to more accurately predict or recognize emotions as they exist in the real world, while providing functionality more meaningful as a result.
From an elementary point of view, there is the purely scientific and technological challenge of building an AI system that is capable of empathizing in a natural way. Unlike systems that only provide information, artificial intelligence carries out operational tasks or makes predictions based on data. However, an empathologist could also help create systems designed to positively impact our own emotions.
In Banissy’s words, “As a researcher in social cognition and empathy, one of the big questions in my field is that social health is a crucial characteristic in life. We know, for example, that loneliness can have very negative effects on health and well-being, so developing a system that can somehow facilitate or provide social benefits for health would be incredibly beneficial. “
Is this, that machines can have empathy, the beginning of that moment that some scientists call superintelligence? Or what is the same, will it be the beginning of a non-biological intelligence superior to the human one?
