Japan 2018

Japanese and South Korean imaginings of robots and intelligent machines are frequently evoked as an alternative to Western depictions. But to what extent are they really different, and how? How do East Asian imaginings shape the public perception and technological advancement of AI in those areas? On 12 September 2018, we held the second Global AI Narratives Workshop at Waseda University Tokyo to explore these questions.

Professor Toshie Takahashi, from the School of Culture, Media and Societyat Waseda University, had organised the workshop in collaboration with CFI, the Research Institute for Letters, Arts and Sciences at Waseda, The Japanese Society for Artificial Intelligence and The Robotics Society of Japan. It was opened by the Vice President of Waseda University, Professor emeritus Shuji Hashimoto. He emphasised the importance of narratives at a time when AI was posing many challenging questions, citing Umberto Eco’s The Name of the Rose (1980): “Those things about which we cannot theorise, we must narrate.”

There followed four sessions. First, the AI Narratives team from CFI – Dr Kanta Dihal, Dr Stephen Cave, and Dr Sarah Dillon – presented some of their work so far: Dr Dihal on the influence of Asimov’s laws of robotics on science policy; Dr Cave on the predominant hopes and fears for AI in the Anglophone West; and Dr Dillon on her ‘What AI Researchers Read’ project. The question and answer session brought out a point that recurred throughout the day: while in Japan many AI researchers are influenced by comics and anime, the role of the graphic in Western Anglophone narratives is much smaller, and the most influential ones (the Marvel and DC comics) only tangentially deal with AI.

The next panel featured four Japanese researchers, starting with Professor Takahashi herself. Her survey of young people in Japan brought out fascinating perspectives: for example, that 73% of Japanese youths view AI positively, while at the same time, 60% do not want robots to look humanoid. Her work has taken her to the oldest as well as the youngest citizens of Japan: also featured on the panel was the famous robot seal PARO, whose maker, Professor Takanori Shibata, was present at the workshop. In nursing homes, the introduction of PARO reduced the tension between residents and nurses. The role of assistive technology for elderly-care services, a pressing issue in ageing Japan, was also touched upon in the presentation of Dr Kentaro Watanabe. Japan, as well as South Korea, as we learned later, is facing a serious social problem in a society that is ageing at an extremely high rate. AI and robots therefore have significant potential for social good in these societies.

Masayoshi Sakai returned to graphic genres in his presentation on AI and robots in manga and animé. He introduced the two concepts according to which Japanese AI and robot characters can be categorized: the ‘buddy’ and the ‘extension’. Astroboy and Doraemon are the two ‘buddy’ characters that have most strongly shaped Japanese conceptions of AI and robots. Respectively in the form of a child and a cat, they are also notably different from the two most influential Western equivalents, Robby the Robot and the Terminator. The ‘extension’ characters, which have been around in manga and anime since at least 1956 (Tetsujin-28 or ‘Iron Man 28’), are notable for explicitly being (mere) tools: Tetsujin is not autonomous, but a servant of good or evil depending on who possesses his remote control.

Professor Osamu Sakura from the University of Tokyo touched upon another major theme in comparing Eastern and Western AI narratives: while the West juxtaposes humans to both non-human animals and to AI/robots, Eastern narratives emphasise continuity between these categories. This is a result of the different philosophical and religious traditions prevalent in Japan, which, he argued, have been more influenced by Taoism and relatively unaffected by the Abrahamic religions.

The South Korean panel followed. It had an equal gender balance, an approach the CFI team actively facilitates throughout our Global AI Narratives project. Professor Kyung Sin Park noted that this balance is unfortunately not common in South Korea itself, which he said is one of the most unequal societies for women to live in. Nonetheless, the Korean panel had a strong emphasis on feminism, including Dr Kyoungmi Oh presenting on the history of computing as a profession for women.

Professor So-Young Kim emphasised first and foremost the shock South Korea experienced when in 2016 the British AlphaGo system defeated their Go champion Lee Sedol (a theme picked up by other panellists). The media used the phrase ‘Robot Spring’, reminiscent of the ‘Arab Spring’. A large amount of funding was subsequently poured into AI and robotics research in Korea. The government framed this as a catch-up strategy, an attempt to reduce a perceived gap between Korea and more advanced countries. She explained that AI was communicated to the public following familiar narratives that had been used for emerging technologies in the past.

Professor Chihyoung Jeon explained that the problem of an ageing society is equally pressing in South Korea. Therefore, in both Japan and South Korea, the older generation in particular does not consider robots a threat to jobs as in Western countries: they are not replacing people – they are being introduced in areas where there is not enough human labour. However, Professor Jeon showed several examples of reality TV shows in which human-robot interactions failed. These shows explored the effects of introducing humanoid robots in elderly people’s lives, including a woman who was given a robot ‘grandchild’ and a man given a robot drinking buddy; the woman’s quality of life was improved much more through interaction with the production team.
After the three panels, the workshop attendees were divided into groups for a discussion session. One group consisted of Professor Takahashi’s students, who had been immensely helpful throughout the day, to discuss what young people in Japan thought of AI. With the average student in the group being born in 1997, their perspectives surprised some other Japanese attendees. First, they found that their attitudes towards AI were divided by gender, with the men in the group having a more positive view of AI, and the women a more negative one. Second, the students stated that manga and anime such as Astroboy and Doraemon had not influenced them as much as it had earlier generations.

Doraemon, nonetheless, was used as the default picture for illustrating AI stories in newspapers, another discussion group found. This was a striking contrast to the Terminator picture so prevalent in the UK. Nonetheless, another discussion group noted that the Terminator was not entirely absent: in Korea, this image was used to illustrate more cautious narratives about AI, although these were seen as external to South Korea’s techno-utopian outlook.

Professor Takahashi and the CFI team are working on a more detailed report of this meeting, and we look forward to building on the many links it forged. In the meantime, this is a summary of the clearest points to emerge. Japanese and Korean narratives around AI are:

-       much less concerned about automation related job loss than those of the UK and US;
-       much less concerned with themes of robot rebellion;
-       tend to portray intelligent machines as friends, helpers or (reliable) extensionsof humans, instead of as murderous, rebellious Others;
-       influenced by a different philosophical tradition, which emphasises connectedness between kinds of entity (such as human, animal, machine);
-       concerned with keeping up with, or catching up with, other countries (like the UK and US) that they perceive to be ahead.