Anxiety Sensitive Artificial Intelligence (AnxSAI) mimics, simulates, replicates, recognizes, interprets, and responds to human anxiety, with the intention of enhancing the system’s ability to interact with humans in more meaningful ways or of providing realistic support for human decision.
AnxSAI systems can be categorized as
Human-centered AnxSAI: sensing and adapting to (factors of) anxiety.
Human-like AnxSAI: replicating (human) anxiety and related behaviors.
AnxSAI research
AnxSAI is a highly transdisciplinary question, involving at once the psychological understanding of anxiety and how to model it as well as the many benefits that anxiety can offer for AI research and the ethical, social, and cultural ramifications of deploying AnxSAI systems. Key questions involve:
can we use AI systems to better understand the processes of anxiety?
can we make more human-friendly AI systems by making them sensitive to the anxiety their decisions can create?
can we help therapists and patients to better deal with anxiety-sensitive disorders?
can we improve AI systems by incorporating properties of human anxiety in their behavior?
can we use AI systems to simulate the manifestations of human-like anxiety, for training and policy-making purposes?
Key AnxSAI publications
An overview of AnxSAI and potential research tracks:
Horned, A., & Vanhée, L. (2023, May). Models of Anxiety for Agent Deliberation: The Benefits of Anxiety-Sensitive Agents. In Proceedings of the 2023 International Conference on Autonomous Agents and Multiagent Systems (pp. 1761-1767). URL
A review of human-like AnxSAI for social simulation:
Horned, A. and Vanhée, L.; (2022) From Threatening Pasts to Hopeful Futures. A review of agent-based models of anxiety. In Social Simulation Conference; pp. (publishing in progress) URL
Technical implementations of AnxSAI are available:
Gutsche, L. and Vanhée, L.; (2023) The Value of Knowledge: Joining Reward and Epistemic Certainty Optimisation for Anxiety-Sensitive Planning (best workshop paper). In Workshop on Interdisciplinary Design of Emotion-sensitive Agents; pp. (publishing in progress); Publisher: Springer
Vanhée, L. and Jeanpierre, L. and Mouaddib, A.-I.; (2022) Anxiety-Sensitive Planning: From Formal Foundations to Algorithms and Applications. In Proceedings of the International Conference on Automated Planning and Scheduling; pp. 730–740; URL
Extensive work relating anxiety and IT technologies is broadly available, including master theses carried along the AnxSAI's umbrella.
If you find your research suited for AnxSAI, please contact us.