The Amplifying Effect of Explainability in AI-assisted Decision-making in GroupsIn the era of artificial intelligence, AI-assisted decision-making has become a common paradigm. Explainable Artificial Intelligence has been one of the more explored factors in improving transparency of AI tools in AI-assisted decision-making, but sometimes with contradictory results. Furthermore, while individual AI-assisted decision-making has garnered substantial investigation, the domain of group AI-assisted decision-making remains notably underexplored. This research presents the first look at the impact of explainability and team composition on AI-assisted decision-making. With a controlled experiment on mushroom edibility classification, with 89 participants, we show that the impact of XAI is more pronounced in decision-making with groups (2-person) than in individual decision-making. Groups rely less on incorrect AI recommendations when explanations are available, but they rely more on incorrect AI recommendations when explanations are absent, compared to individual decision makers. This phenomenon underscores the amplified effect of explainability in AI-assisted decision-making in group settings.2025RDRegina de Brito Duarte et al.INESC-ID; University of Lisbon, Instituto Superior TécnicoExplainable AI (XAI)AI-Assisted Decision-Making & AutomationCHI
Coding Together: On Co-located and Remote Collaboration between Children with Mixed-Visual AbilitiesCollaborative coding environments foster learning, social skills, computational thinking training, and supportive relationships. In the context of inclusive education, these environments have the potential to promote inclusive learning activities for children with mixed-visual abilities. However, there is limited research focusing on remote collaborative environments, despite the opportunity to design new modes of access and control of content to promote more equitable learning experiences. We investigated the tradeoffs between remote and co-located collaboration through a tangible coding kit. We asked ten pairs of mixed-visual ability children to collaborate in an interdependent and asymmetric coding game. We contribute insights on six dimensions - effectiveness, computational thinking, accessibility, communication, cooperation, and engagement - and reflect on differences, challenges, and advantages between collaborative settings related to communication, workspace awareness, and computational thinking training. Lastly, we discuss design opportunities of tangibles, audio, roles, and tasks to create inclusive learning activities in remote and co-located settings.2023FRFilipa Rocha et al.Faculdade de Ciências, Universidade de LisboaCognitive Impairment & Neurodiversity (Autism, ADHD, Dyslexia)Programming Education & Computational ThinkingSpecial Education TechnologyCHI
I Know I Am, But What Are You? How Culture and Self-Categorization Affect Emotions Toward RobotsAs routinely working with robots spreads globally, it becomes important to understand how best to customize robots to each culture to maximize collaboration between humans and robots. In two distinct cultures (United States and Portugal) we examined group-based emotions toward robots with participants self-categorizing three different ways (ingroup, outgroup, and neutral). We tested and confirmed our baseline assumptions that Portuguese participants are more collectivistic and less individualistic, and feel closer with a team in negative, but not positive, scenarios, compared to United States participants. Supporting our hypotheses, the results showed that participants rated more positive emotions toward the robot in the ingroup condition than in the outgroup or neutral conditions. Moreover, an interaction effect between culture and self-categorization revealed that Portuguese participants had more positive group-based emotions toward the robot than United States participants when self-categorizing as an ingroup. We discuss the implications in terms of human-robot teaming and potential future research directions.2022FCFilipa Correia et al.Human-Robot Interaction; Human-Robot InteractionCSCW
Community Based Robot Design for Classrooms with Mixed Visual Abilities ChildrenVisually impaired children (VI) face challenges in collaborative learning in classrooms. Robots have the potential to support inclusive classroom experiences by leveraging their physicality, bespoke social behaviors, sensors, and multimodal feedback. However, the design of social robots for mixed-visual abilities classrooms remains mostly unexplored. This paper presents a four-month-long community-based design process where we engaged with a school community. We provide insights into the barriers experienced by children and how social robots can address them. We also report on a participatory design activity with mixed-visual abilities children, highlighting the expected roles, attitudes, and physical characteristics of robots. Findings contextualize social robots within inclusive classroom settings as a holistic solution that can interact anywhere when needed and suggest a broader view of inclusion beyond disability. These include children's personality traits, technology access, and mastery of school subjects. We finish by providing reflections on the community-based design process.2021INIsabel Neto et al.University of LisbonUniversal & Inclusive DesignSpecial Education TechnologySocial Robot InteractionCHI
Creativity Encounters between Children and RobotsCreativity is an intrinsic human ability with multiple benefits across the lifespan. Despite its importance, societies not always are well equipped with contexts for creativity stimulation; as a consequence, a major decline in creative abilities occurs at the age of 7 years old. We investigated the effectiveness of using a robotic system named YOLO as an intervention tool to stimulate creativity in children. During the intervention, children used YOLO as a character for their stories and through the interaction with the robot, creative abilities were stimulated. Our study (n = 62) included 3 experimental conditions: i) YOLO displayed behaviors based on creativity techniques; ii) YOLO displayed behaviors based on creativity techniques plus social behaviors; iii) YOLO was turned off, not displaying any behaviors. We measured children’s creative abilities at pre- and post-testing and their creative process through behavior analysis. Results showed that the interaction with YOLO contributed to higher creativity levels in children, specifically contributing to the generation of more original ideas during story creation. This study shows the potential of using social robots as tools to empower intrinsic human abilities, such as the ability to be creative.2020PAPatrícia Alves-Oliveira et al.STEM Education & Science CommunicationSocial Robot InteractionHRI
Investigating the Opportunities for Technologies to Enhance QoL with Stroke Survivors and their FamiliesThere are over 80 million stroke survivors globally, making it the main cause of long-term disability worldwide. Not only do the challenges associated with stroke affect the quality of life (QoL) of survivors, but also of their families. To explore these challenges and define design opportunities for technologies to improve the QoL of both stakeholders, we conducted semi-structured interviews with 10 survivors and one of their family members. We uncovered three major interlinked themes: strategies to cope with technological barriers, the (in)adequacy of assistive technologies, and limitations of the rehabilitation process. Findings highlight multiple design opportunities, including the need for meaningful patient-centered tools and methods to improve rehabilitation effectiveness, emotion-aware computing for family emotional support, and re-thinking the nature of assistive technologies to consider the perception of transitory stroke-related disabilities. We thus argue for a new class of dual-purpose technologies that fit survivors' abilities while promoting the regain of function.2020ISInês Santos Silva et al.INESC-ID & Instituto Superior Técnico, Universidade de LisboaVibrotactile Feedback & Skin StimulationMental Health Apps & Online Support CommunitiesElderly Care & Dementia SupportCHI
Warping Deixis: Distorting Gestures to Enhance CollaborationWhen engaged in communication, people often rely on pointing gestures to refer to out-of-reach content. However, observers frequently misinterpret the target of a pointing gesture. Previous research suggests that to perform a pointing gesture, people place the index finger on or close to a line connecting the eye to the referent, while observers interpret pointing gestures by extrapolating the referent using a vector defined by the arm and index finger. In this paper we present Warping Deixis, a novel approach to improving the perception of pointing gestures and facilitate communication in collaborative Extended Reality environments. By warping the virtual representation of the pointing individual, we are able to match the pointing expression to the observer's perception. We evaluated our approach in a co-located side by side virtual reality scenario. Results suggest that our approach is effective in improving the interpretation of pointing gestures in shared virtual environments.2019MSMaurício Sousa et al.INESC-ID Lisboa & Universidade de LisboaHand Gesture RecognitionSocial & Collaborative VRCHI
Hybrid-Brailler: Combining Physical and Gestural Interaction for Mobile Braille Input and EditingBraille input enables fast nonvisual entry speeds on mobile touchscreen devices. Yet, the lack of tactile cues commonly results in typing errors, which are hard to correct. We propose Hybrid-Brailler, an input solution that combines physical and gestural interaction to provide fast and accurate Braille input. We use the back of the device for physical chorded input while freeing the touchscreen for gestural interaction. Gestures are used in editing operations, such as caret movement, text selection, and clipboard control, enhancing the overall text entry experience. We conducted two user studies to assess both input and editing performance. Results show that Hybrid-Brailler supports fast entry rates as its virtual counterpart, while significantly increasing input accuracy. Regarding editing performance, when compared with the mainstream technique, Hybrid-Brailler shows performance benefits of 21% in speed and increased editing accuracy. We finish with lessons learned for designing future nonvisual input and editing techniques.2018DTDaniel Trindade et al.INESC-ID, Instituto Superior Técnico, University of LisbonVisual Impairment Technologies (Screen Readers, Tactile Graphics, Braille)Augmentative & Alternative Communication (AAC)CHI