Co-designing a Social Robot for Newcomer Children’s Cultural and Language Learning DOI: 10.1145/3776734.3794350 ISBN: 979-8-4007-2321-6/2026/03 Conference: Companion Proceedings of the 21st ACM/IEEE International Conference on Human-Robot Interaction; March 16–19, 2026; Edinburgh, Scotland, UK Companion Proceedings of the 21st ACM/IEEE International Conference on Human-Robot Interaction (HRI Companion ’26), March 16–19, 2026, Edinburgh, Scotland, UK CCS: Human-centered computing Empirical studies in HCI CCS: Human-centered computing Empirical studies in interaction design

Neil Fernandes, Tehniyat Shahbaz, Emily Davies-Robinson, Yue Hu, Kerstin Dautenhahn · 2026

Read Co-designing a Social Robot for Newcomer Children’s Cultural and Language Learning DOI: 10.1145/3776734.3794350 ISBN: 979-8-4007-2321-6/2026/03 Conference: Companion Proceedings of the 21st ACM/IEEE International Conference on Human-Robot Interaction; March 16–19, 2026; Edinburgh, Scotland, UK Companion Proceedings of the 21st ACM/IEEE International Conference on Human-Robot Interaction (HRI Companion ’26), March 16–19, 2026, Edinburgh, Scotland, UK CCS: Human-centered computing Empirical studies in HCI CCS: Human-centered computing Empirical studies in interaction design by Neil Fernandes, Tehniyat Shahbaz, Emily Davies-Robinson, Yue Hu, Kerstin Dautenhahn (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-26674.txt.

Introduction

Fernandes and co-authors report a co-design study for a social robot that helps newcomer children learn language and culture. Literacy programs for these children often face thin staffing, mixed skill levels, and short contact time. The team works with tutors and coordinators to shape a robot named Maple. They map four recurring challenges in the setting and weigh the robot's role in cultural orientation and belonging. Early design guidelines for classroom use close the study, ahead of trials with children and families.

Catalog record

CREATORNeil Fernandes, Tehniyat Shahbaz, Emily Davies-Robinson, Yue Hu, Kerstin Dautenhahn
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE21065
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.26674

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Co-designing a Social Robot for Newcomer Children’s Cultural and Language Learning DOI: 10.1145/3776734.3794350 ISBN: 979-8-4007-2321-6/2026/03 Conference: Companion Proceedings of the 21st ACM/IEEE International Conference on Human-Robot Interaction; March 16–19, 2026; Edinburgh, Scotland, UK Companion Proceedings of the 21st ACM/IEEE International Conference on Human-Robot Interaction (HRI Companion ’26), March 16–19, 2026, Edinburgh, Scotland, UK CCS: Human-centered computing Empirical studies in HCI CCS: Human-centered computing Empirical studies in interaction design Neil Fernandes, Tehniyat Shahbaz, Emily Davies-Robinson, Yue Hu, Kerstin Dautenhahn Abstract . Newcomer children face barriers in acquiring the host country’s language and literacy programs are often constrained by limited staffing, mixed-proficiency cohorts, and short contact time. While Socially Assistive Robots (SARs) show promise in education, their use in these socio-emotionally sensitive settings remains underexplored. This research presents a co-design study with program tutors and coordinators, to explore the design space for a social robot, Maple . We contribute (1) a domain summary outlining four recurring challenges, (2) a discussion on cultural orientation and community belonging with robots, (3) an expert-grounded discussion of the perceived role of an SAR in cultural and language learning, and (4) preliminary design guidelines for integrating an SAR into a classroom. These expert-grounded insights lay the foundation for iterative design and evaluation with newcomer children and their families. 1. Introduction Newcomer children often need to acquire the language of their host country while simultaneously adjusting to unfamiliar school systems, cultural norms, and social expectations ( 30 ) . In this paper, we use the term ‘newcomer children’ to refer to young children from immigrant or refugee backgrounds who have migrated to Canada within the past five years and are in the process of learning English as an additional language, adapting prior definitions of newcomer youth and students used in Canadian education research ( 17 ; 14 ) . Community based literacy programs provide after school support through worksheets and small group activities. Tutors in these settings must support culturally and linguistically diverse groups where children differ in age, first language, country of origin, and proficiency. As a result, they must balance language and socio-emotional support as well as classroom management, often without information about each learner’s proficiency. SARs have been proposed as one way to support learning in such constrained environments. Within Robot-Assisted Language Learning (RALL), social robots have shown to increase motivation and confidence ( 18 ) , support speaking practice ( 20 ) , and scaffold early literacy skills, outperforming virtual agents or tablets ( 19 ) . Social robots are often perceived as more affiliative and less judgmental than screen based systems ( 6 ) , which may be particularly valuable for learners who are apprehensive about making mistakes in front of adults or peers ( 21 ; 13 ) . Work that explicitly considers newcomer populations is beginning to emerge ( 23 ; 22 ; 30 ) , but existing studies are typically short term, rarely situated in community programs. This paper addresses this gap using a co-design approach with tutors and coordinators from United for Literacy (UFL) - a national charitable literacy organization that delivers programs to children, youth and adults in urban, rural and remote communities across Canada - who have direct insight into learners’ needs and the constraints they face as tutors ( 27 ) . Using the Design Thinking process ( 5 ) and a co-design approach ( 15 ; 12 ; 1 ; 25 ) , we conducted a co-design meeting wherein experts reflected on their everyday practice and reacted to a robotic prototype that we call Maple . This paper presents four primary contributions: (1) a domain summary that describes four recurring challenges in supporting newcomer children, (2) a discussion of how cultural orientation and community belonging emerge in experts’ accounts, (3) an analysis of the perceived roles of an SAR in this context, and (4) preliminary design guidelines for a knowledgeable peer-like robot that addresses these challenges while aligning with current pedagogies. 2. Methods and Procedures 2.1. Co-designers Our design team comprised four participants: two expert second-language (L2) tutors 1 1 1 co-authors in paper from UFL that delivers literacy programs 2 2 2 https://www.unitedforliteracy.ca/Programs and two Human–Robot Interaction (HRI) researchers (one professor and one graduate student). One L2 expert was a regional manager overseeing the organization’s programs in the Region of Waterloo with seven years of experience, and the other L2 expert was a Program Support Staff in the region who has been with UFL for two and a half years as both staff and volunteer. 2.2. Robotic System Overview Our prototype, Maple (see Figure 1 ), is assembled from the ROBOTIS Engineer Kit 1 3 3 3 https://en.robotis.com/model/page.php?co_id=prd_engineerkit , previously evaluated in HRI work ( 30 ) . The platform uses multiple DYNAMIXEL servos to obtain expressive upper-body gestures and head movements, and is configured as a compact table-top humanoid (to support coordinated verbal/non-verbal interaction). The robot is adapted to mount a smartphone as the face display, allowing non-verbal gestures to be combined with animated eyes and a moving mouth for social cues. Figure 1 . The Maple robot prototype. The control stack is implemented in Python on the Robotics Operating System (ROS) Noetic Ninjemys distribution, using a modular architecture with separate components for motion control, audio output, facial animation, and higher-level behavior tree ( 7 ) . Facial animation is provided by PyLips ( 9 ) , which synchronizes mouth movements and expressions with robot speech, and the overall scheduling of speech and motion follows principles from RoboSync ( 28 ) . To make interaction flows easy to create and modify, we provide a web-based user interface to be used on a tablet, built in React, similar to ( 16 ) that connects to the robot via a ROS bridge ( 8 ) , allowing tutors and researchers to define story-based activities and quizzes without changing the underlying control code. 2.3. Procedure and Analysis We conducted one 60-minute online co-design meeting with the participants as a group over Microsoft Teams. The meeting followed a semi-structured, interview-style format and was recorded and transcribed. In the first part of the session, experts reflected on their everyday practice and current challenges in the community language learning program. In the second part, the experts watched a demo of a suggested interaction with the working Maple prototype and brainstormed possible activities based on the robot’s capabilities. The data was then summarized using the domain summary method ( 4 ) , which resulted in the themes that capture recurring challenges and pedagogical goals, which formed the basis for the preliminary guidelines. 3. Preliminary Results 3.1. Challenges Across the first co-design meeting, experts described four recurring challenges: (1) Language Barriers and Skill Levels, (2) Lack of Motivation and Attention., (3) Absence of Formal Proficiency Baselines, and (4) Lack of One-on-One Attention and Support. 3.1.1. Language Barriers and Skill Levels Language barriers were consistently identified as an immediate difficulty. The recruitment of learners is very inclusive, and before sessions start tutors will not have prior knowledge of the children’s individual language proficiencies. Learners are highly diverse, originating from many language backgrounds with variable English proficiency. In some cases, children “ […] do not speak a single word of English […] ,”

Read the complete text