Is Japan's AI 'the second battleship Yamato'?
A few days ago, the Japanese AI research institute, 'Dwango Artificial Intelligence Laboratory' has been closed due to its financial deficit. Dwango used to be in the forefront of AI research in Japan, but its history is very short --only four and half years.
In 2014, I attended the summer session of the Japanese Cognitive Science Society. The them of the session was 'Brain science and AI'. The head of Dwango was invited to the session as the main speaker. After his lecture, many famous university professors and grad students crowded around him and asked questions about his research. I noticed no one in this session was interested in my specialty, Naturalistic Decision Making (NDM), at all.
In 2017, I published my first book about human expert vs. AI. In my book, I emphasized the point that 'Because human intuitive thinking is quite different from algorithmic thinking of AI, AI cannot simulate human creativity and problem solving exactly' 'AI cannot surpass human completely.'
In fact, Dr. Gary Klein kindly shared his personal research experience at the US Military Laboratory with me. His team and he had a project to develop AI that could think like humans. Although they tried to input human thought processes through programming several times, the computer screen showed 'ERROR!' After all, they gave up their project. As a matter of fact, I heard from Klein and other related people that the US military forces do not depend on AI. There are risks such as malfunction and hacking.
Developing human-like AI looks like building the Japanese battleship Yamato. The former Japanese imperial Navy built Yamato to prepare for the war against the US. However, the US shifted its tactic from so-called 'fleet-to-fleet decisive battle' to 'aircraft carrier battle'. However, the Japanese Navy hanged on fleet-to-fleet decisive battle due to success during the Russo-Japanese War. The result: Japan wasted the cost of Yamato and thousands of Marines and lost the war.
One of the most crucial aspects of research on NDM and human factors and ergonomics would be that we study how we can promote 'Human-centered development' of machines, not 'Machine-centered development'. As the Japanese Navy commanders and engineers of Yamato neglected human factors and ergonomics, present Japanese still don't consider the relationship between human and machines. Therefore, I call Japanese AI the second battleship Yamato.
Indeed, history repeats itself.