AI is the intelligence displayed by machines, as opposed to the natural intelligence displayed by humans and animals with awareness and emotion. The difference between the first and last category is usually clear from the acronym you choose.
The best information processing can be expected by combining humans and technology specialties.
So far, we have accumulated a lot of knowledge in various fields such as psychology, behavioral science. And cognitive science as academic fields to study the ideal way of human resources and organizations. Although it is a new research field compared to them, “brain science” is currently attracting attention.
By scientifically investigating the function of the brain. Which controls human decision-making and behavior. It is expected that productivity will be improved, innovation will be realized. And business will be significantly transformed.
Professor Yoshikuni Edakawa, a professor at Waseda University’s Research Strategy Center, specializes in the themes of brain science, corporate management, organization, and human resources. And makes various proposals for the front lines of business.
This time, we asked Professor Edakawa about “HR technology from the perspective of brain science.” What does HR technology represented by AI (artificial intelligence) mean to humans, and how should we use it? It is packed with valuable tips for solving inevitable problems for future human resources.
Nowadays, there is increasing interest in applications to corporate management. And human resources fields.
In terms of relationships with personnel and organizations disciplines. Such as industrial psychology, organizational psychology, behavioral science. And cognitive science has once explained human behavior. There, people and the brain are a kind of “black box”, and it was common to guess the contents of the black box by investigating a lot of what kind of input and what kind of output.
However, in brain science, the function of the brain itself, which is regarded as the black box, will be investigated by experiments using the above-mentioned fMRI and the like. The big difference is that we can explain in a straightforward manner. Is what kind of information processing mechanism works in the brain for input. And the thoughts and actions as output come out.
Since brain science is still a new discipline. There are some parts that can explain human behavior and some that cannot. However, I have gradually come to explain the key parts.
What is happening in the brain when communicating between people? What is the function of the brain to make decisions? If we can understand how the brain is involved in these behaviors. That are carried out every day in the field of business. We can expect “the reproducibility of the law”. Which is one of the conditions for being scientific. It can be explained that “people tend to behave in this way. Because the brain works like this under these conditions,” and it is possible to make predictions.
Human behavior is not simple but in situations such as business and management where the speed of change is extremely fast and it is difficult to grasp the target, from the brain that can scientifically explain the fundamental nature of humans. The approach is rather effective. I think that kind of expectation is rising.
“Deep learning” incorporates the learning mechanism of the brain
In the area of HR technology, AI (artificial intelligence) is drawing attention. What is the relationship between AI deep learning and brain science?
Today’s AI is said to be the “third generation,” but the key is the technology called “deep learning.” In the past, “machine learning” was performed, but it is to give AI the information prepared in advance by a person who plays the role of a teacher to learn. It was, so to speak, passive learning. Deep learning, on the other hand, is closer to the process of human “intelligence” development in that AI learns spontaneously.
In fact, deep learning was developed based on the mechanism by which the human brain learns. As humans grow from babies to adults, they learn many things in their daily lives besides studying at school. Deep learning allows AI to voluntarily do something similar to this kind of experiential learning. If you show 10,000 pictures of cats, AI will be able to grasp the characteristics and recognize “this is a cat” when the next “cat-like thing” comes. In particular, AI today is evolving steadily in terms of such image recognition.
It means that elucidating the function of the brain led to deep learning technology.
However, deep learning is not perfect yet. I said that it is modeled on the function of the brain, but there are many paths in the human brain to establish learning, and there is a question as to whether all of them can be incorporated into AI. If you expect it to be a substitute for human intelligence, it will not be fully realized if there are parameters that are overlooked. There is also the question of whether the current brain science research itself has been able to elucidate all of the phenomena of “learning.” I think there is also the idea of proceeding with research and development by thinking that “it’s okay to some extent”, but at least you should be aware that the current AI and deep learning are still developing.
The use of HR technology, including AI, is expanding. How do you see this situation?
The beauty of technology is that it efficiently handles “work that people cannot do” and “work that is difficult to do.” I think it’s basically a good thing that the utilization progresses.
There is often a debate about AI. Such as “whether it is mastered or used”. And although there are serious discussions among experts and experts. The current AI is a person in a specific area, for example, image processing. It is a stage that is better than. Since such AI is still “weak AI”. It moves according to its own judgment.
Such as “strong AI” that is installed in the anime Tetsuwan Atom and Doraemon, that is, general-purpose AI. Then, when AI can communicate with people is realized. There may be a problem close to the so-called “singularity” of how to coexist. However, at this stage, I think it should be greatly utilized as a “tool” that can perform a fixed range of work with more than human ability.
People’s “intuition” is often “pattern recognition”
It is also argued that by utilizing HR technology. Human resources should shift from analog “experience and intuition” to digital “data and evidence”.
The characteristic of human decision-making and information processing is that it utilizes “heuristics”. Heuristics are human-like judgments that the accuracy of the results is coarse. But the time it takes for the results to be derived is short. So even if the input information is limited, you can make a convincing judgment to some extent.
You can give it down. However, since early AI was “finding the optimal solution from all combinations”. It was possible to deal with vague problems such as “what to eat for lunch today” in a short time. did not. However, humans can quickly decide based on heuristics, “I had curry yesterday. So let’s make ramen today.”
When talking about this, it is easy to understand it as an episode that “AI is still a long way off”. But intuitional judgments by heuristics are often based on “pattern recognition”. Experienced managers can make accurate decisions because they know various patterns and recognize patterns at a level below their level of consciousness. Since this pattern recognition is a field where technologies such as AI are good. It is fully expected that it will be close to human judgment in the future.
However, even in the field of human resources. It is not a matter of immediately switching from “experience and intuition” to “data and evidence.” In response to the voice, “I want you to find such a human resource”. The work of selecting from a large number of candidates based on the specifications is exactly patterned recognition. To it is faster to use technology and unnecessary bias can be eliminated. One thing to keep in mind is whether the parameters that make up the pattern are appropriate. People may feel “somehow good” because they are aware of invisible parameters. If we can take into account such invisible parameters, it will be possible to process information more efficiently using technologies such as AI.
Does it mean that people should do the fields where the parameters cannot be exhausted?
The program that runs these technologies must include a number of parameters. Parameters that exist at the time of programming can be included, but those that do not exist cannot be included. In other words, the parts that have not been clarified by brain science at this point cannot be implemented in AI. One of the characteristics of human beings is that they have a part that is invisible to the eye and cannot be written. But that they “certainly feel”. I think that is a major factor that AI cannot completely replace people today.