Unknown risks and the collapse of human civilisation: A review of the AI-related scenarios

Abstract

Science and technology have experienced a great transition, a development that has shaped all of humanity. As progress continues, we face major global threats and unknown existential risks even though humankind remains uncertain about how likely unknown risks are to occur. This paper addresses five straightforward questions: (1) How can we best understand the concept of (existential) risks within the broader framework of known and unknown? (2) Are unknown risks worth focusing on? (3) What is already known and unknown about AI-related risks? (4) Can a super-AI collapse our civilisation? Furthermore, (5) how can we deal with AI-related risks that are currently unknown? The paper argues that it is of high priority that more research work be done in the area of 'unknown risks' in order to manage potentially unsafe scientific innovations. The paper finally concludes with the plea for public funding, planning and raising a general awareness that the far-reaching future is in our own hands.