Trust in Artificial Intelligence: The case of rehabilitation robots (Japanese)

The relationship between robots and humans is becoming increasingly similar to relationships between humans. As shown in preceding studies, trust between humans and robots has been discussed for a long time, but at this point, there are no concrete guidelines on what exactly constitutes trust, and the issue is still under debate. This paper discusses trust in using rehabilitation robots which are medical devices whose use and development involves multiple parties. Based on the four-way relationship between doctors, caregivers, robot developers, and patients, we report the results of a questionnaire survey that analyzed the elements of trustworthy AI from the standpoint of each party. In conclusion, there are differences in the elements that comprise the definition of a “trustworthy robot†which are deemed desirable by the four different categories of parties. For example, engineers do not demand the “low cost†that patients, doctors, and caregivers demand. There are also differences in what elements are seen as desirable by each party at the time of an epidemic, such as COVID-19, and during normal times. We point out that Japan needs to fill in these discrepancies through its policies for future technological development of rehabilitation robots, which will be indispensable in compensating for the shortage of human resources in the medical field as the decline in the birthrate and the aging of the population continue unabated.