Towards a computational model for higher orders of Theory of Mind in social agents.

Abstract

Effective communication between humans and machines requires artificial tools to adopt a human-like social perspective. The Theory of Mind (ToM) enables understanding and predicting mental states and behaviours, crucial for social interactions from childhood through adulthood. Artificial agents with ToM skills can better coordinate actions, such as in warehouses or healthcare. Incorporating ToM in AI systems can revolutionise our interactions with intelligent machines. This proposal emphasises the current focus on first-order ToM models in the literature and investigates the potential of creating a computational model for higher-order ToM.