I Know I Should: Normative Competence From Biology To AI

How do people understand what is normatively expected from them? And how does normative cognition motivate action? A distinct feature of human sociality is our capacity to sense the appropriate thing to say in a certain context or the action that should be done and to care enough to do it. Such competence with social norms stabilizes joint action between dyads and enables cooperation among thousands, but whether it is the product of domain-general learning mechanisms or dedicated cognitive ones is still unclear (Heyes, 2024). Moreover, as the many blunders made by contemporary AI systems like LLMs show, autonomous AI agents too need to acquire some level of normative competence to be reliable partners. However whether this is possible with current architectures is still debated (Browning, 2024). The aim of this symposium is to bring together researchers to discuss the biological, cognitive, computational and interactive bases of normative motivation and judgment.