TRI-TECH FINAL WORKSHOP
ASIMOV: Adaptive Social Interaction based on user’s Mental mOdels and behaVior in HRI
September 10, 2025
International Conference on Social Robotics + AI
Centro Direzionale, Isola C4, Room 6 – Floor 1°, 14:30 – 18:30
Organizers: Salvatore M. Anzalone, Giuseppe Palestra, Carlotta Bettencourt, Elena Ricci, Alberto Pirni, Maryam Alimardani, Mariacarla Staffa
Program
Paolo Dario - Invited speaker talk: “The history and the future of robot companionship”
Filippo Cavallo - Contributed talk: TBD
Alberto Pirni - Contributed talk: “Ethically Rethinking Human-Robot Interaction in the Framework of Elderly Society: the TRI-TECH PROJECT”
Q&A
Coffée break
Diego Reforgiato, Daniele Riboni, Lorenzo Boi, Federica Casula, Silvia M. Massa - Contributed talk “LLM-Powered Companion Robots in Elderly Care: Technical and Ethical Perspectives”
Diego Reforgiato, Daniele Riboni, Lorenzo Boi, Federica Casula, Silvia M. Massa - Contributed talk “Exploring Ethical and Technical Dimensions of Companion Robots in Elderly Society: The TRI-TECH Experience in Cagliari”
Elena Ricci - Contributed talk “Exploring Ethical and Technical Dimensions of Companion Robots in Elderly Society: The TRI-TECH Experience in Pisa”
Q&A
Triniti Armstrong, Ross Sowell, William D. Smart, and Naomi T. Fitter - Contributed paper “Explaining Robots with Edutainment”
Hesham M. Shehata and Mohsen Farid - Contributed paper “Advancing Mental Health Support with AI Social Robots: Scenario-Driven Communication Using Large Language Models”
Kristiina Jokinen and Graham Wilcock - Contributed paper “Agents of Conversation: Grounding and Cognitive Architectures”
Domenico Palmisano, Giuseppe Palestra, Berardina Nadja De Carolis - Contributed paper “Enhancing Social Robots through Resilient AI”
Giovanna Varni - Contributed paper: TBD
Marlena R. Fraune, Giovanna Varni, Alberto Pirni - Panel: “Multi-Party Human-Robot Interaction in Vulnerable Contexts: Mental Models, Trust, and Ethics”
Abstracts
The history and the future of robot companionship
Paolo Dario
TBD
Filippo Cavallo
TBD
Ethically Rethinking Human-Robot Interaction in the Framework of Elderly Society: the TRI-TECH PROJECT
Alberto Pirni
TBD
LLM-Powered Companion Robots in Elderly Care: Technical and Ethical Perspectives
Diego Reforgiato, Daniele Riboni, Lorenzo Boi, Federica Casula, Silvia M. Massa
The TRI-TECH project introduces a robotic platform that integrates humanoid robots (NAO and Pepper) with Large Language Models to enhance elderly care. The system architecture combines speech-to-text, intent recognition, dynamic response generation, and multimodal interaction, enabling natural and adaptive communication. Key technical challenges addressed include latency optimization, modular scalability, and seamless integration of gestures, speech, and contextual feedback. The platform aims to test how advanced AI can be embedded into social robots to deliver reliable, customizable, and human-like interaction. Development has been guided not only by technical requirements but also by ethical considerations, ensuring responsible and human-centered innovation in elderly care.
Exploring Ethical and Technical Dimensions of Companion Robots in Elderly Society: The TRI-TECH Experience in Cagliari
Diego Reforgiato, Daniele Riboni, Lorenzo Boi, Federica Casula, Silvia M. Massa
In Cagliari, the TRI-TECH project conducted an experimental study with 15 participants over the age of 70 to evaluate their interaction with humanoid companion robots. Eligibility was determined through a neuropsychological assessment: MoCA (cognitive status) and ADL/IADL scales (autonomy in abilities). Then, interaction sessions between the elderly and the robots were conducted, supervised by the research team. Each participant underwent a semi-structured pre-use and post-use interview, complemented by standardized tests. This methodology enabled direct comparison of pre- and post-interaction perceptions, providing insights into acceptance, trust, attitudes and emotional response towards companion robots.
Exploring Ethical and Technical Dimensions of Companion Robots in Elderly Society: The TRI-TECH Experience in Pisa
Elena Ricci
In Pisa, the TRI-TECH project conducted an experimental study with 15 participants over the age of 70 to evaluate their interaction with humanoid companion robots. Eligibility was determined through a neuropsychological assessment: MoCA (cognitive status) and ADL/IADL scales (autonomy in abilities). Then, interaction sessions between the elderly and the robots were conducted, supervised by the research team. Each participant underwent a semi-structured pre-use and post-use interview, complemented by standardized tests. This methodology enabled direct comparison of pre- and post-interaction perceptions, providing insights into acceptance, trust, attitudes and emotional response towards companion robots.
Explaining Robots with Edutainment
Triniti Armstrong, Ross Sowell, William D. Smart, and Naomi T. Fitter
The integration of modern robots, from food delivery robots to self-driving cars, into our daily lives is pervasive and exciting. At the same time, there is a distinct lack of understanding of robotic technologies among the general population, including workers who use these systems and policymakers who govern them. In this paper, we present an approach that interested parties can use to facilitate edutainment based activities meant to help general populations understand the basic functioning and capabilities of robots. We conducted early assessments of the effectiveness of these intervention experiences during experiments in a conference setting (N = 18) and a classroom setting (N = 13). The results suggest that participants who self-identified as less experienced in robotics tended to feel more informed about robotic technologies after experiencing the perspective-taking activities (compared to their more experienced counterparts). Others who are interested in public engage ment, robotics education, and policy-making for new technologies can benefit from the presented work.
Advancing Mental Health Support with AI Social Robots: Scenario-Driven Communication Using Large Language Models
Hesham M. Shehata and Mohsen Farid
Artificial intelligence (AI) technologies—particularly large language models (LLMs)—are increasingly being investigated for mental health applications. This study explores how AI social robots integrated with LLM can engage users in scenario-based therapeutic conversations tailored to mental health conditions such as anxiety, depression, ASD, ADHD, and Alzheimer’s disease. By combining scenario design with conversational AI and social robotics, we evaluate how these technologies can support emotional engagement and trust. Using case studies informed by structured prompts and real-time interaction, we assess LLM-generated dialogues in terms of user receptiveness, empathy, and feedback and clinical and psychological accuracy. Results should show promise for nonjudgmental, accessible support while also raising essential considerations around bias, privacy, and clinical accuracy. This ongoing research work contributes to the responsible deployment of LLM-based AI social robots in future mental health care.
Agents of Conversation: Grounding and Cognitive Architectures
Kristiina Jokinen and Graham Wilcock
Moving from simple LLM-based chatbots to complex agentic workflows has recently caused renewed interest in cognitive architectures. More sophisticated agentic frameworks provide more advanced options for both short-term and long-term memory. This short paper describes one of the new frameworks (LangGraph) and explores how it might be used to represent common ground in human-robot interactions.
Enhancing Social Robots through Resilient AI
Domenico Palmisano, Giuseppe Palestra, Berardina Nadja De Carolis
As artificial intelligence continues to advance and becomes more integrated into sensitive areas like healthcare, education, and everyday life, it’s crucial for these systems to be both resilient and robust. This paper shows how resilience is a fundamental characteristic of social robots, which, through it, ensure trust in the robot itself—an essential element especially when operating in contexts with elderly people, who often have low trust in these systems. Resilience is therefore the ability to operate under adverse or stressful conditions, even when degraded or weakened, while maintaining essential operational capabilities.
Giovanna Varni
TBD