2025_programme: Advancing a Passive Acoustic Simulation to Research Empathetic Artificial Intelligence (EAI) in Human-AI Teaming

Back


  • Day: June 17, Tuesday
      Location / Time: D. CHLOE at 17:40-18:00
  • Last minutes changes: -
  • Session: 06. Enhancing underwater acoustic sensing through machine learning
    Organiser(s): Cheng Chi
    Chairperson(s): Peng Xiao
  • Lecture: Advancing a Passive Acoustic Simulation to Research Empathetic Artificial Intelligence (EAI) in Human-AI Teaming
    Paper ID: 2287
    Author(s): Madalin Facino, Alan Hunter, Christof Lutteroth, Aaron Roberts, Samantha Dugelay
    Presenter: Madalin Facino
    Abstract: This paper reports advancements in the development of a serious game environment designed to investigate empathetic artificial intelligence (EAI) within the domain of human sonar operations. Building on previous foundational work, the project incorporates significant upgrades, including the integration of a more advanced transmission loss model, procedural path generation for entities informed by established behavioral frameworks, and extensive enhancements to the user interface (UI). These refinements ensure the simulation upholds ecological validity while offering a more immersive and methodologically rigorous platform for psychological research.\n\nThe serious game simulates the high-stress environment of sonar operations, where operators analyze complex acoustic signals under cognitive and emotional strain. Key improvements focus on representing realistic underwater acoustic phenomena while balancing cognitive load and perceptive difficulty to evoke measurable psychological responses. Procedural generation of entity behaviors enhances the adaptability and realism of operational scenarios, creating dynamic challenges for participants.\n\nThe concept of empathy is currently being examined through a systematic literature review aimed at defining EAI behaviors. Insights from this review will guide the implementation of adaptive AI support systems capable of responding to real-time operator states. Experimental investigations will compare the impact of EAI to standard AI on operator trust, cognitive load, and task performance under varying conditions of operational demand and AI reliability.\n\nThis interdisciplinary effort bridges computer science, psychology, underwater acoustics, and engineering, aiming to advance human-AI teaming in sonar operations. By leveraging psycho-physically informed game mechanics and advanced acoustic simulations, the serious game establishes a robust platform for exploring how EAI can mitigate the psychological challenges inherent in underwater acoustic tasks.\n
  • Corresponding author: Ms Madalin Facino
    Affiliation: University of Bath
    Country: United Kingdom

Back