Artificial intelligence -A transformative force: Facing the challenges ahead!

Abstract

Artificial Intelligence (AI) is considered to be a transformative force that is bound to alter the fabric of society. The recent major advances in machine learning are revealing AI´s capacity as a general-purpose technology and push inventions in areas of mobility, healthcare, home & service robotics, education and cyber security, to name just a few. AI-enabled developments have promising capabilities to increase human well-being and to resolve, inter alia, the grand challenges associated with for instance our ageing society or climate change. At the same time, AI comes with risks and challenges associated to fundamental human rights, ethical issues and broader societal implications. I will sketch both, capabilities and risks related to AI and discuss implications from a policy perspective:I will start my talk with providing a definition for AI systems and some conceptual clarifications relating to buzzwords like "digitalization" or "big data" found in popular literature. Economists have 6classified AI systems as "prediction machines" (Agrawal et al, 2018) which have the capacity to produce precise and reliable for-and nowcasts to reduce uncertainty for decision makers. It is this capacity that makes AI a game changer that is getting economically exploited: to better deal with uncertainty, enabled through enormous amounts of various kinds of data and data processing technologies.By taking a decision theory perspective, I will explore the role of AI technology in prediction, judgement and reasoning processes, all contributing to decision making. This analysis of an Input -Process -Output (IPO) Decision Model allows a systematization of potential challenges associated with AI systems. In more detail, I will discuss issues with (i) data quality and data governance, (ii) goal specification and goal conflicts (iii) judgement and prediction models, (iv) integration of artificial and human agents into socio-technical systems, and finally (v) broader socio-cultural implications of AI deployment. In order to provide concrete examples, I will discuss some AI applications like predictive analytic modeling, moral decision problems of autonomous vehicles, and implications for work and labor markets.I will end my talk with a discussion of implications for further research and policy recommendations. I conclude that there is a need for ethical guidelines beyond regulatory frameworks to ensure that AI systems increase human well-being without doing harm.