ABSTRACT
Artificial intelligence (AI) systems are increasingly integrated, often invisibly so, in our physical and online environments transforming our society in a sociotechnical system. We use AI to replace, support and extend human cognitive abilities. There is a lot of promise, but also there is a lot of hype in evaluating the abilities and the role of AI in the future. On one hand, we do not fully understand the processes that we are automating such as image recognition, behaviour prediction, empathy. On the other hand, AI executed cognitive tasks are not the same process that is human cognition, making it necessary to pay careful attention. For both reasons, the ethical impact AI automation has on our lives and society is difficult to evaluate. To understand the threats and prospects of AI, we need to understand its limits: in what ways will AI outperform us as a tool and with which consequences?
