JAE INTRO - 6 months internship at the Institut de Robótica e Informática Industrial
Duration: 7 months, 20 hours per week.
Funding: 4.200 € in total, financed 100% by CSIC.
Title: Toward Robots that Understand Us: Theory of Mind and Socially Aware Decision-Making
Proposal: Recent advances in Artificial Intelligence have led to highly capable systems, yet current AI still struggles to understand people, social context, and the norms that guide human behaviour. Socially intelligent robots and agents must be able not only to perceive the environment but also to reason about users’ intentions, adapt to social expectations, and select appropriate actions in dynamic situations. Enabling this level of social awareness remains a central challenge for next-generation AI.
This project investigates how socially aware AI systems can be designed by integrating Theory of Mind (ToM), social norm reasoning, and hybrid neuro-symbolic approaches. Theory of Mind enables artificial agents to infer others’ intentions, beliefs, and goals, supporting proactive and cooperative behaviour. At the same time, social norms provide behavioural constraints that shape acceptable actions through mechanisms analogous to reward and punishment. Combining these dimensions allows AI systems to move from purely reactive behaviour toward context-sensitive and socially aligned decision-making.
The internship will explore computational models that allow robots and intelligent agents to: (1) infer user intentions from multimodal interaction signals, (2) represent and reason about social norms across contexts, and (3) adapt their behaviour through learning from feedback and interaction histories. Particular attention will be given to integrating symbolic knowledge representation with data-driven learning methods and possibly implementing that on our robots: Temi, TIAGo, Stretch, and Ari. At IRI we have a new living Lab called Labora 650mq2, with a real flat of 65m2 in which it is possible to test your implementation in a real setting.
Training Plan:
The student will acquire interdisciplinary training at the intersection of Artificial Intelligence, cognitive modelling, and human-robot interaction. The training plan includes:
1. Scientific background acquisition
- Literature review on Theory of Mind in AI, social norms modelling, and socially aware robotics.
- Introduction to neuro-symbolic AI and hybrid decision-making architectures.
2. Technical skills development
- Fundamentals of reinforcement learning and probabilistic modelling for intent recognition.
- Knowledge representation techniques (ontologies, knowledge graphs, structured memory). - Use of modern AI tools and foundation models for interaction modelling.
3. Research methodology
- Problem formulation, experimental design, and evaluation of socially aware behaviours. - Development of prototype components.
- Reproducible research practices and collaborative development.
4. Transversal skills
- Scientific writing and presentation of results.
- Participation in group meetings, seminars, and research discussions.
- Exposure to interdisciplinary research linking AI, cognition, and social interaction.
By the end of the internship, students are expected to understand the main challenges of socially intelligent AI, implement initial experimental prototypes, and gain experience in conducting supervised research within a multidisciplinary environment. Upon completion of the internship, there will be opportunities to continue working with us on your TFG, TFM, or PhD.
How to apply: Candidates must submit their applications by April 11th 2026, on this website: Requisits The call is addressed to students who are in any of the following situations:
- To be enrolled in a Bachelor's (Grado) in the academic year 2025-2026,
- - If the Grado has 240 ECTS, at least 180 ECTS of their credits must be completed,
- - If the Grado has 300 ECTS or is a Double Grado, at least 240 ECTS of their credits must be completed
- Or to be enrolled in a Master's in the academic year 2025-2026 or 2026-2027
- Not having previously received another JAE grant.
- Average GPA for the Bachelor's, in 0-10 scale, should be equal to or greater than 7.5