This ZiF long-term group workshop explores links between decision-making under uncertainty in natural systems and the design of robust AI systems. The group focuses on two directions: designing robust and transparent AI systems by drawing on models of decision-making in humans, insects, and other biological systems; and using AI systems driven by decision heuristics to better understand decision-making in natural systems.

Decision heuristics build on Herbert Simon’s idea of bounded rationality. They use adaptive satisficing to make rational decisions “in the wild”, where the space of possible scenarios is not known in advance. Examples include tallying, sequential search, fast and frugal trees, and algorithmic methods for identifying heuristics suited to a domain and task.

The workshop continues the group’s work on robustness, decision-making over a time horizon, explainability, and the question of when decision heuristics should be used alongside other mathematical formulations for decision-making. Participants will work with experts in human behavior sciences, risk modeling, bio-robotics, computational neuroethology, and complex systems theory.

Convenors: Mohan Sridharan (Edinburgh), Özgür Şimşek (Bath), Konstantinos Katsikopoulos (Southampton), and Gerd Gigerenzer (Berlin).

Full event details are available on the ZiF event page.