REACT Logo

REACT

REsponsible and Adaptive robots in Complex seTtings

About the Project

REACT is a Coordinating Keystone sub-project within RAI UK (Responsible AI UK), with a dedicated focus on responsible AI development and deployment. We aim to develop responsible and adaptive robotic systems that can operate in complex social and healthcare environments, ensuring that our intelligent systems understand context, adapt to changing conditions, and maintain the highest ethical standards in human-robot interactions.

As part of the RAI UK initiative, this multidisciplinary project addresses the increasing demand for care in ageing societies and the shortage of available human caregivers. By developing multimodal multi-agent robotic systems that can understand and adapt to human needs, we aim to enhance the quality and efficiency of care without replacing essential human interaction. Our work exemplifies the responsible AI principles that guide the broader RAI UK mission.

Research

Our research explores how multi-human multi-robot (MHMR) adaptive systems can be responsibly integrated into complex social environments, particularly within health and social care. We investigate how robots can learn from human behavior, adapt to social contexts, and maintain safety in unpredictable environments.

The project is structured into five interconnected work packages (WPs), with co-creation embedded throughout all stages. Our multidisciplinary approach combines co-design with stakeholders, real-world scenario prototyping, physiological and behavioural sensing, and explainable adaptive AI. Work Package 3 (Socio-Technical Co-Production and Implementation) serves as the central hub, informing and supporting all other work packages through continuous stakeholder engagement.

Overview of REACT Project Work Packages

Figure 1: Overview of all Work Packages in the REACT project

Work Packages:

WP1: Modelling, Preference Aggregation
Design multiagent models and scenarios, multimodal preference aggregation, and ensure responsible robot behavior.
WP2: Multimodal Dataset Creation
Deploy sensors to understand humans, integrate physiological and behavioural sensors, and build a comprehensive multimodal dataset.
WP3: Socio-Technical Co-Production
Central coordinating work package. Engage stakeholders for real-world needs, work with stakeholders to identify needs, and inform design and ethical considerations.
WP4: AI Training and Deployment
Train AI models for adaptive behavior, recognize user preferences in real-time, and enable adaptive behaviors in robots.
WP5: System Evaluation and Validation
Evaluate system performance and ethics, assess transparency, accountability, and compliance, and ensure responsible deployment in healthcare.

We work closely with diverse stakeholders—including older adults, carers, healthcare professionals, care home residents, technologists, designers, and ethicists—to ensure the technologies we develop are informed by real-world experiences, needs, and values. This collaborative approach ensures that our robotic systems are context-aware, emotionally empathetic, socially intelligent, and ethically grounded.

Datasets

Access datasets generated through our research, available for open use under appropriate ethical guidelines. Our data contributes to the global research community working on responsible AI and adaptive robotics systems.

The project will produce multimodal datasets that integrate neurophysiological and behavioural data, along with AI models that interpret user preferences, emotional states, and care-related requirements.

Publications

Discover our latest research findings and publications in leading conferences and journals. Our work contributes to the understanding of responsible AI, adaptive robotics, and human-robot interaction in complex social environments.

We publish our findings in top-tier venues covering robotics, AI ethics, human-computer interaction, and healthcare technology. Our publications demonstrate the practical applications and theoretical foundations of our adaptive robotic systems.

Team Members

Meet the multidisciplinary team behind REACT. Our researchers bring expertise from robotics, artificial intelligence, human-computer interaction, psychology, and ethics to create comprehensive solutions for complex real-world challenges.

Aleksandra Landowska
School of Computer Science, University of Nottingham
aleksandra.landowska@nottingham.ac.uk
Aislinn D Gomez Bergin
School of Computer Science, University of Nottingham
aislinn.bergin@nottingham.ac.uk
Ayodeji O. Abioye
School of Computing and Communications, The Open University
ayodeji.abioye@open.ac.uk
Jayati Deshmukh
School of Electronics and Computer Science, University of Southampton
j.deshmukh@soton.ac.uk
Andriana Boudouraki
School of Computer Science, University of Nottingham
andriana.boudouraki@nottingham.ac.uk
Maria Waheed
School of Computer Science, University of Nottingham
maria.waheed@nottingham.ac.uk
Athina Georgara
School of Electronics and Computer Science, University of Southampton
a.georgara@soton.ac.uk
Dominic Price
School of Computer Science, University of Nottingham
dominic.price@nottingham.ac.uk
Tuyen Nguyen
School of Electronics and Computer Science, University of Southampton
tuyen.nguyen@soton.ac.uk
Shuang Ao
Responsible AI, University of Southampton
s.ao@soton.ac.uk
Lokesh Singh
School of Electronics and Computer Science, University of Southampton
lb5e23@soton.ac.uk
Yi Dong
University of Liverpool
yi.dong@liverpool.ac.uk
Rafael Mestre
School of Electronics and Computer Science, University of Southampton
r.mestre@soton.ac.uk
Joel E. Fischer
School of Computer Science, University of Nottingham
joel.fischer@nottingham.ac.uk
Sarvapali D. Ramchurn
School of Electronics and Computer Science, University of Southampton
sdr1@soton.ac.uk

Contact Us

Get in touch with our project coordinators for more information about REACT and our research.

👤 Aleksandra Landowska
👤 Ayodeji O. Abioye