SWATE: IA Socialmente Inteligente para suporte e treino de trabalho em equipa

SWATE aims to investigate and address the significant challenges in unpacking complex human behaviour in teamwork processes. Specifically, explores how low-level data traces from commodity sensors can inform teamwork high-level constructs and serve as evidence of the team and each element’s capabilities. This would be a set towards creating technology that improves teamwork efficiency and efficacy. We outline the following scientific and technological challenges: [G1] SWATE will advance the state-of-the-art by identifying key indicators of effective teamwork, using continuous and unobtrusive capture from cameras placed in high-fidelity environments. These indicators hold the potential to forecast team (and individual) outcomes and offer valuable insights for both trainers and trainees into their performance and the team's adaptive coordination. Utilizing cameras in conjunction with speech signals, we intend to decipher high-level social activities and their underlying intentions, from low-level data traces. This goal also addresses the importance of capturing the temporal aspect of data, which is often overlooked. [G2] Develop an AI with advanced knowledge of environmental dynamics at the individual and team levels. SWATE will develop a set of new algorithms that enable the conceptualization of team performance in terms of interpersonal and cognitive processes. The former focuses on action and transition phases between actions as situation monitoring, communication and assistance on task-specific processes. The latter is concerned with how the team collects and processes information to build a collective understanding or a shared mental model. This goal deals with the development of the “agent mind”, the mechanism that combines knowledge from different disciplines in order to be able to judge what the team knows about the task, what the team needs to do to accomplish it and assess teamwork competencies according to its core components. [G3] SWATE will enhance team training through AI support and actionable insight. SWATE will design and implement a collaborative interaction between the AI and the team trainer, to augment the trainer's perception of social processes and coordinative adaptations. The project intends to furnish trainers with practical tools for designing complex training interventions, provide richer and in-context feedback to trainees and support debrief reflections. [G4] Investigate the application of the developed tool in healthcare teams, particularly in clinical simulation scenarios, using GSLMED state-of-the-art facilities. Clinical simulations provide a controlled yet realistic environment for healthcare professionals to practice various scenarios, from routine procedures to high-stress emergencies, without endangering patients' lives. This includes prioritizing fidelity, and ensuring that the simulation accurately mirrors real-world situations, equipment, and procedures, thereby promoting an authentic learning experience. Our goal is to create a training environment that not only enhances technical skills but also prepares healthcare professionals for the psychological demands and complexities they may face in their practice. This approach ultimately aims to improve patient outcomes by ensuring that healthcare teams are well-prepared and resilient in diverse clinical scenarios.