Implicit Values in Generative AI: What Does AI "Want"?

This lecture explores the ethical implications and underlying values of generative AI systems as part of the 'AI and Values' seminar series. Join VUB professor and CLEA research director Francis Heylighen as he examines what drives AI systems, whether they possess inherent goals or values, and how these values manifest. Delve into the tension between AI's effectiveness in problem-solving and concerns about manipulation, bias, and potential human control. Heylighen, whose research focuses on complex systems, self-organization, and collective intelligence, approaches these questions through an ontology of actions and processes rather than material entities. Learn how systems emerge as self-maintaining networks of processes and how this perspective illuminates our understanding of AI behavior. This 93-minute presentation was delivered at the VUB university campus in Brussels on March 14, 2025, as the inaugural session of the AI and Values Seminar series, which is open to both in-person and online participants.