Autonomy- Oriented Justification of Norms for Trustworthy AI Systems in the Age of Global Digitalization

Some assume that the rise of risky AI systems is the end of the Enlightenment and its related basis for trust, the autonomy-oriented justification and reasoning of norms. A first look at utopias and dystopias in innovations seems to confirm that. However, a second look on the legal framework suggests the opposite. The legal system offers some established and autonomy-oriented normative standards, that have proven their trustworthiness already in many societal fields. These standards are too often set aside in the regulation of AI systems. Nevertheless, they would significantly contribute to an autonomy-oriented legal- systematic classification of AI systems. Based on that, the relationship between risky AI systems and trustworthy normative justifications is discussed with examples of current regulatory proposals. I argue for avoiding the "mechanistic trap" in the ongoing juridification of AI systems in favor of a more autonomyoriented approach.