Social Impacts of Artificial Intelligence and Mitigation Recommendations: An Exploratory Study

Abstract

With all the changes and adaptations of the industrial sector to migrate to a more technological and connected environment, society also needs to adapt to a new format of work and lifestyle. Therefore, the challenge of evaluating the social impact related to Artificial Intelligence (AI) becomes evident, since the speed of this transformation leads society to rethink the man-machine interaction, as well as the impacts related to the new forms of work that emerge every day. From this perspective, this study aimed to develop a systematic review of the literature and evaluate qualitatively the main social impacts linked to the use of artificial intelligence, as well as the main mitigation actions, using as database the Web Of Science and Scopus. Nine social impacts and twelve mitigation actions were mapped. Among the impacts, bias and discrimination growth due to the use of AI (26%) stand out, followed by the possibility of autonomous systems hurting humans (16%), and then the impact related to the violation of data, privacy, and freedom (15%). It was evident that the recommendation considered to be the most comprehensive refers to increasing the breadth and depth of public debate regarding the risks of AI application. In second place is the recommendation to create laws, principles, and regulations for the use of AI. The results of this study point to the protection of life and human ethics, which are topics that need to be widely investigated for the dissemination of knowledge, and to enable the sustainable development of AI.