Impacts of AI on human and society are discussed on the basis of concrete understanding of current technology and its problems. First, an overview of the basics of AI technology (neural network, deep learning, knowledge based system) and its current status are presented, followed by a view on the wave of the future (real world and humanity). Then, the problems with current AI (reliability/safety, explainability, biases, value misalignment, difficulties with real world and humanity) are discussed, together with current status of frontier research for future development. Finally, discussions on the future of AI and human/society are presented on two focuses; (1) Should AI be endowed with human-like mind? and (2) What role AI should play in our future society.

　 Some states are now focusing in research and development of autonomous weapon using Artificial Intelligence (AI), which is expected by them as critical tools to advantageously expand the war. However there are accusations by some other states for its inhumanity. United Nations has started discussion officially on this issue last November.

　In Japan, Ministry of Defense (MOD) is focusing military application of advanced technology including AI autonomous weapon, and started research funding for academic researchers. In response, there has been debate in Japan on whether academic researchers should commit military research or not. In this paper, current international situations and facts on AI autonomous weapons, and debate on social responsibility of scientist in Japan are examined.

　 In this article, we aimed to examine what kind of robots are accepted in Japanese society as more and more human robots are being developed. We outlined the study on psychological ANSHIN for robots and reported that “Comfort” and “Human-likeliness” on robots were found as factors as ANSHIN. Regarding cognitive attribution of human essence, there are two aspects of cognitive function learned in society and richness of emotions that human beings have in nature. In comparison with the Westerners, Japanese tend to distinguish between positive and negative of these two aspects. We discussed the necessity to consider the cultural values of Japanese society and the experience of contact in time series.

　 This paper introduces preceding studies which question human-machine interface and interaction from the perspective of artificial intelligence. Studies introduced in this paper are mainly books from several disciplines and practical technologies. Three disciplines (ethnomethodology, cognitive science and human-agent interaction) and two practical technologies (communication robot and autonomous driving) are included in the studies. Each discipline and technology has different approach, but they all emphasize situated relationship between human and machine which involves external environment, not only paying attention to the internal model of the subject. Also, in the research of practical technologies, discussions about social and ethical issues have been made from the perspective of human-machine interaction.

　 Technology assessment is undertaking of assessing social implications of technology and conveying the results of assessment to stakeholders for supporting their decision and policy making. This article picks up major potential issues of technology assessment relating to artificial intelligence (AI) and reviews institutional options of technology assessment, that is, international technology assessment, technology assessment by research institutes, setting up and utilization of fund by government for technology assessment implemented by various institutes, technology assessment by national governments (legislative branch and administrative branch). After reviewing the attempts of technology assessment in general and experiments of technology assessment on AI by various institutions, one experimental attempts by IICP (Institute for Information and Communications Policy) by MIC (Ministry of Internal Affairs and Communication) of Japan is analyzed. This experiment by IICP has characteristics in that technology assessment is undertaken with some distance from the R＆D and the actual utilization of technology and that target of technology assessment is rather comprehensive compared to other attempts of technology assessment by other institutions. The issues of relationship between developers and users (providers of data set for learning) and the issues of transition when users of AI and non-users of AI coexist are identified. Finally, actual and potential impact of AI on political decision making is also briefly analyzed, as one aspect of social implications.

　These days, IT-skill accompanied with “creativity” has been expected to wider public. In this research, we examined whether the creativity of “programming education” can be visualized using “task dividing workshop”. We analyzed the result of task dividing workshop for junior high school students, participants in the Student Session of Japanese Society for Artificial Intelligence(JSAI), and graduate students. We analyzed the results of the task division work from three points: ①characteristics of how tasks are increased, ②characteristics of classification of tasks by group, and ③reading of sentences. we also found that the ability to “Extending tasks from experience” of junior high school students was less than other groups. Data implies that all groups members have similar ability to divide task into subtasks. From the analysis of task work, we visualize the creativity that each group have. By using these findings, we clarified the viewpoint of evaluating programming thinking that uses creativity, and that it can be expected to improve learners’ abilities. By using these findings and visualizing creativity necessary for programming thinking, we suggested that it lead to students’ subjective learning.

　 “Today we are at the beginning of a fourth industrial revolution. It is characterized by much more ubiquitous and mobile internet, by smaller and more powerful sensors that have become cheaper, and by artificial intelligence and machine learning” （Schwab 2017, 3）. For example, Japanese companies are starting to use AI in recruiting and financing. It is said that The Kanagawa police plan to launch AI-based predictive policing system.

　 In United States, several states have already put a predictive policing system in place. Moreover, other states have adopted risk assessment tools in sentencing. COMPAS is the most popular risk assessment tool. However, it is not disclosed that how COMPAS determines individual risk scores or it how weighs various factors in arriving at a risk score. Critics pointed out that COMPAS's algorithm is biased against minorities. In this kind of situation, Wisconsin Supreme Court allowed state to continue using COMPAS to assist in sentencing （State v. Loomis, 881 N. W. 2d 749）.

　 This paper considers how we should constitutionally use AI from Loomis's case study.

　 It is pointed out that “information technology” can be effective in human intellectual work while also having a negative influence on the user’s mental state. On the other hand, the field of psychology is examining Wellbeing index to show a good state of mind and attempts of Positive Computing to implement this in information technology are beginning. In this paper, as a method of information technology methodology to promote Wellbeing, we examined a methodology focused on human unconsciousness.

　 The society is getting more globalized and complex with development of information technology. Accordingly, people are getting to have each diversified values, and to be required to understand it for preventing social troubles. We research what is needed to be aware of values, and aim to construct a system with the algorism. In this paper, we introduce a prototype system that support aware of values and idea generation (AIR-VAS). We speculate that knowing other community’s opinions in the discussion, is connected to the new idea generation and awareness of diversified values. We developed the system on the approach that visualizing discussion information as word frequency and word co-occurrence network, and we used the system in a workshop. Based on the result, we analyzed an association between discussion and visualized information, and evaluated an effectiveness of the system.

　 This paper provides four comments on social and multi-stakeholder discussions in Japan about ELSI of robots and artificial intelligence technologies. These comments are 1) information overload, 2) specific issues of robots and AIs, 3) legacies of information ethics and computer ethics, 4) values. The aim of this paper is to contribute to future discussions by exploring these issues.

　 The expectation for research on dual-use technology was urged by the defense-technology diplomatic policy to reduce procurement costs and to acquire defense equipment technologies (especially next-generation fighter aircraft); moreover, it was necessary to improve bargaining power in anticipation of international joint development in the future. In addition, by the diplomatic policy in science and technology to improve Japan’s soft power and the innovation-oriented policy in science and technology under industry-university-government cooperation, it was expected that universities would engage in dual-use technology research. In the discussion of a diplomatic strategy for science and technology, its regulation and ethics were excluded from the agenda by the cabinet office. The acquisition of globalized, advanced, and high-priced conventional arms, and spin-on/off dual-use technology was the focus. As affirmative support for dual-use technology research was carried on regardless of the political parties in power, it seems a certain agreement was reached among Japanese politicians or was caused by some other external influence.