Full Professor of Artificial Intelligence at Instituto Superior Técnico, University of Lisbon, and founder of GAIPS at INESC-ID. I build social agents and robots that perceive, adapt, collaborate and act socially with people.
For more than three decades my research has centred on one question: how do we build autonomous agents and robots that interact with humans in a fair, transparent, social and natural way?
I am a Full Professor of Artificial Intelligence in the Department of Computer Engineering at Instituto Superior Técnico (IST), University of Lisbon, and a senior researcher at INESC-ID. In 2000 I founded GAIPS, the Group on AI for People and Society, one of the internationally leading groups in social AI, which I coordinated for twenty years.
My work engineers AI systems that exhibit specific social capabilities: social perception, social adaptation, following norms, collaboration, resolving social conflicts, and verbal and non-verbal behaviour, among others. I bring these together in a framework for Artificial Social Intelligence, and apply them across intelligent learning environments, social robotics, social simulation, serious games and entertainment. Throughout, I place the human in the loop, working towards AI that is trustworthy and centred on human values: AI for people and society.
I also believe that AI is changing the way we live, work, and, for us scientists, the way we teach and conduct science. Our society will change and new policies must be developed to adapt to this new world. Between April 2024 and June 2025 I served as Secretary of State for Science in the XXIV Portuguese Government, overseeing national science and technology policy, and contributing to this world in change.
There is a difference between social artificial intelligence and Artificial Social Intelligence. My work is about the latter: giving machines the intelligence to act socially, so they can perceive, adapt, and take part in the social worlds of people. A single framework of social capabilities ties the threads of my research together.
To interact well with people, an agent needs more than task competence. It needs to read a social situation and act within it. I study how to engineer these capabilities and how to compose them, so that agents and robots become genuine social partners rather than mere tools. The framework anchors every thread that follows.
The core framework: engineering social understanding, interpersonal competencies, social decision making and social responsibility into agents, so that they can act socially. This thread relates to, and underpins, all the others.
Affective computing and empathy in agents and robots: modelling emotions, recognising them, and using them in the communication between humans and machines, including long-term interaction and the role of humour in human-robot interaction.
Human-robot interaction beyond the one-robot, one-user paradigm: robots that take part in groups and teams, hybrid societies of humans and robots, membership and team formation, trust, and inclusion.
Engineering pro-sociality with autonomous agents: cooperation, fairness and social dilemmas studied through multi-agent systems and evolutionary game theory, towards societies of humans and agents that are more cooperative and altruistic.
Intelligent learning environments and social good: empathic tutors, education for children with neurodevelopmental conditions, serious games, storytelling, cultural understanding and social simulation.
XXIV Portuguese Government, 2024 to 2025
AAAI Fellow since 2023. EurAI and ELLIS Fellow
Principal Investigator on 27 national and international projects, attracting more than €7M in funding for the group. A selection follows.
One of my major tasks is supporting the next generation of researchers in social robots and Artificial Social Intelligence. I have mentored 4 post-docs and supervised 27 PhD students. My former students now lead research and teaching across the world. At the moment I supervise two PhD students (together with Prof. Joana Campos).
More than 80 course instances over the years in Artificial Intelligence, human-computer interaction and computer science. Four courses were designed from scratch: Autonomous Agents and Multi-Agent Systems, Affective Computing, Computer-Based Learning, and Social Robotics and Human-Robot Interaction.
More than 100 invited and keynote talks, from neuroscience and the social sciences to autonomous agents and robotics. A recent selection.
Over 300 articles in journals, books and the proceedings of high-impact conferences, accumulating more than 21,000 citations. A curated selection follows; the full record is on Google Scholar and ORCID.