AnxSAI: Anxiety-Sensitive Artificial Intelligence

Project description

Anxiety is an inherent component of human experience. As a mental state, anxiety creates the mental tension and physical readiness for dealing with possible threats and uncertainties --anxiety is argued to be an activator for our cognitive machinery to be active when facing issues that matters. Yet, anxiety is also a harm, known to be a primary inducer for risky behaviors and mental health disorders (depression, post-traumatic stress disorders, etc), inducing yearly social costs estimated to outreach dozens of trillions of SEK (trillions euros).

Artificial Intelligence (AI) and anxiety are tightly related and significant scientific and social value can be obtained from connecting them with each other. The AnxSAI initiative is dedicated to study all the possible interplays between artificial intelligence and anxiety:

• 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 understanding anxiety be useful for making better AI systems?
• can we use AI systems to simulate the manifestations of human-like anxiety, for training and policy-making purposes?

See updated AnxSAI webpage here. 

Project Members 

• Melania Borit, Professor in Knowledge Integration and Interdisciplinary Science, modelling and simulation of social processes & game-based learning, UiT the Arctic University of Norway
  • Has offered methodological support for various lines of AnxSAI research and strategic insights on establishing the field
• João Carvalho, intern in the Responsible AI team and student at Umeå University.
  • Has completed a MSc thesis on anxiety and online dating (MSc thesis), cum laude (VG grade)
• Anthony Eid, student at Umeå University, Sweden
  • Has completed a MSc thesis on anxiety and absentminded scrolling (MSc thesis)
• Linda Gutche, student at Ecole Normale Superieure d'Ulm, France
  • Has completed a MSc thesis on models of epistemic anxiety (on embargo until publication), cum laude (VG grade, 18.5/20)
• Arvid Horned, PhD student in the Responsible AI team, Department of Computing Science, Umeå University, Sweden
  • Topic: Psychological foundations of AnxSAI and preliminary models. More information can be found here.

Related institutions

• TAIGA, Umeå's center for transdisciplinary artificial intelligence and TAIGA's social AI focus area

Related activities

• Interdisciplinary Design of Emotion Sensitive Agents (IDEA) - International Workshop co-located with AAMAS 2023, May 2023
• Sense & Sensibility: Modelling human deliberation and decision-making special track at the Social Simulation Conference, September 2023

Funded projects

• 2023: WASP-DDLS funded 500.000 SEK for foundational research on AI & Cognitive Behaviorial Therapy --one of the most scalable approach to dealing with the consequences of exposure to anxiety
• 2024-2028 VR funded 3.800.000 SEK for advanced research on anxiety-sensitive automated planning