Artificial Intelligence, Automation, and Social Welfare: Some Ethical and Historical Perspectives on Technological Overstatement and Hyperbole

摘要

The potential societal impacts of automation using intelligent control and communications technologies have emerged as topics in recent writings and public policy initiatives. Constructed entities labelled as ‘thinking machines’ (such as IBM’s Watson as well as intelligent chatbot and robotic systems) have also played significant roles in this discourse. This paper provides an historical sequencing then analyses a selection of writings produced since the 1940s concerning economic and social issues involving artificial intelligence (AI) research and applications. The paper explores how overstatements and hyperbolic themes and concepts, often stemming from AI’s early periods (including from Herbert Simon), are being employed in characterisations of current AI approaches in apparently opportunistic attempts to provide rhetorical support for various large-scale business and societal initiatives. It also addresses the relative neglect of consideration of many of AI’s sociotechnical failures and discontinued approaches in recent examinations of automation and social welfare issues. The paper discusses the moral logic of AI researchers and developers providing reasonable and measured narratives in public discourse rather than hyperbole, efforts that can empower decision makers to make sounder judgments concerning the technology’s current and future applications as well as allocate rewards of the technology more equitably.