SET-PAiREd : Designing for Parental Involvement in Learning with an AI-Assisted Educational Robot

Hui-Ru Ho, Nitigya Kargeti, Ziqi Liu, Bilge Mutlu · 2025

Read SET-PAiREd : Designing for Parental Involvement in Learning with an AI-Assisted Educational Robot by Hui-Ru Ho, Nitigya Kargeti, Ziqi Liu, Bilge Mutlu (2025) 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-2502-17623.txt.

Introduction

Hui-Ru Ho, Nitigya Kargeti, Ziqi Liu, and Bilge Mutlu design tools that let parents steer a robot tutor for young children. AI learning robots are entering homes, and parents worry about content while welcoming help they lack time or skill to give. Their 2025 work builds SET, a card kit that lays out scenarios with different levels of parent involvement. A companion interface, PAiREd, lets parents review and revise the robot's AI-made lessons before delivery. Twenty families with children aged three to five tried the system at home.

Parents could dial their role up or down, choosing what the robot handles alone. The field study shows how families actually negotiate that control. The authors offer a design pattern for family AI: oversight by design, not by warning label. Their cards make an abstract choice concrete.

Catalog record

CREATORHui-Ru Ho, Nitigya Kargeti, Ziqi Liu, Bilge Mutlu
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE90 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-17623

Browse the full library · About the Tower of Babel

Full text

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

SET-PAiREd : Designing for Parental Involvement in Learning with an AI-Assisted Educational Robot Hui-Ru Ho, Nitigya Kargeti, Ziqi Liu, Bilge Mutlu Abstract . AI-assisted learning companion robots are increasingly used in early education. Many parents express concerns about content appropriateness, while they also value how AI and robots could supplement their limited skill, time, and energy to support their children’s learning. We designed a card-based kit, SET , to systematically capture scenarios that have different extents of parental involvement. We developed a prototype interface, PAiREd , with a learning companion robot to deliver LLM-generated educational content that can be reviewed and revised by parents. Parents can flexibly adjust their involvement in the activity by determining what they want the robot to help with. We conducted an in-home field study involving 20 families with children aged 3–5. Our work contributes to an empirical understanding of the level of support parents with different expectations may need from AI and robots and a prototype that demonstrates an innovative interaction paradigm for flexibly including parents in supporting their children. Conference: CHI Conference on Human Factors in Computing Systems; April 26-May 1, 2025; Yokohama, Japan CHI Conference on Human Factors in Computing Systems (CHI ’25), April 26-May 1, 2025, Yokohama, Japan DOI: 10.1145/3706598.3713330 ISBN: 979-8-4007-1394-1/2025/04 7119 CCS: Human-centered computing Collaborative interaction CCS: Human-centered computing Empirical studies in HCI CCS: Applied computing Interactive learning environments Figure 1. We explored parental involvement scenarios using the SET activity kit, examined parents’ perceptions of AI-generated content, and analyzed their use of PAiREd for collaborating with LLM and robots to support their children’s learning. † † cc-license: by 1. Introduction Parental involvement is essential to children’s early education ( Ma et al., 2016 ; Harris and Goodall, 2008 ) . However, despite parents’ desire to participate in their children’s learning, many of them face challenges, including limited capability, availability, and motivation ( Green et al., 2007 ; Ho et al., 2024 ; Hara and Burke, 1998 ) , which hinder effective participation. For example, one parent may struggle to come up with developmentally appropriate questions to ask their child, while another may know what to discuss but feel unable to dedicate sufficient time due to demanding work-home schedules. AI-assisted robots hold significant potential to transform both the scope and methods of parental involvement in children’s early education. Prior work has highlighted that AI-powered systems can personalize educational experiences to address individual learning needs ( Zhang et al., 2024 ; Xu et al., 2022 ; Xu et al., 2024 ) and that social robots can facilitate rich interactions that support children’s cognitive and emotional development in the physical world ( Ho et al., 2023 ; Leyzberg et al., 2012 ; Belpaeme et al., 2018 ; Kim et al., 2024 ; Lee et al., 2022 ) . Yet, existing studies have primarily focused on addressing child-centric needs, overlooking parental roles. Additionally, parents often raise concerns about privacy, misinformation, and content appropriateness ( Oswald et al., 2020 ; David-Ferdon and Hertz, 2007 ; Howard et al., 2021 ) , emphasizing the need for parental supervision to ensure alignment of AI and robots with parents’ educational goals ( Ho et al., 2024 ; Han et al., 2024 ) . In light of these gaps, it is essential to understand the contexts in which parents support their children’s learning, investigate how AI-assisted robots can be designed to help parents overcome barriers in these contexts, and address concerns regarding children’s interactions with these technologies. We propose that providing parents with flexible control over AI-generated content and collaborative involvement with educational robots can effectively support parents in responding to dynamic contexts and barriers, balancing the benefits and risks of AI-assisted robots. More specifically, this paper explores the following research questions: • RQ1: What contexts do parents encounter when involving in young children’s learning activities? • RQ2: How do parents perceive AI-generated learning content for young children? • RQ3: How would parents prefer to collaborate with LLM on supervising content creation under different contexts? • RQ4: How would parents prefer to collaborate with an AI-assisted robot to engage in learning activities with their children under different contexts? To address these questions, we developed two tools: (1) a card-based activity kit named SET for systematically understanding parental involvement contexts, and (2) a prototype AI-assisted robot named PAiREd to explore flexible parent-AI and parent-robot collaboration mechanisms. The SET kit, with its name representing the key factors of S kill, E nergy, and T ime ( Ho et al., 2024 ; Green et al., 2007 ) , helps parents reflect on their experiences in supporting their children’s learning. Parents use the kit to create scenario cards that capture relatable contexts framed around these factors, serving as both study findings and research tools for contextualizing parent’s interaction with the AI-assisted robot. In addition, the PAiREd prototype ( P arenting with Ai -assisted R obot for Ed ucation) enables parents to generate, edit, and revise storybook content in collaboration with a large language model (LLM). During parent-child reading sessions, parents can adjust task delegation between themselves and the robot via a flexible mechanism, allowing control over the autonomy of the AI-assisted robot. This approach allows parents to balance their involvement based on the specific needs of different scenarios. We evaluated the SET-PAiREd system through a four-stage study. First ( RQ1 ), parents used SET to document real-life involvement scenarios, revealing contextual needs and scenarios. Second ( RQ2 ), interviews about AI-generated content showed that parents value adaptability, efficiency, and affordability, but worry about age appropriateness, inaccuracies, and over-reliance. Third ( RQ3 ), parents co-created learning content for two books using PAiREd , indicating that difficulty, variety, and quality drive their evaluation of LLM-generated content. Under the SET scenarios, their collaboration with the LLM varied depending on their time, energy, trust in the LLM, and interest in editing content. Finally ( RQ4 ), parents read the books with their children, adjusted role delegation between themselves and the robot according to the scenarios, and provided feedback. We identified several collaboration patterns, including using parent takeover mode with LLM-generated content, collaborative modes guided by parent skill and motivation, and robot takeover mode under parental supervision or with reviewed LLM content. Our work makes the following contributions: (1) Prototype Artifact : We developed the SET activity kit to map parents’ real-life contexts around their skill, energy, and time. In addition, we developed and tested PAiREd , a system supporting responsible supervision of LLM-generated content and enabling flexible parent-robot collaboration. (2) Novel Insights : Field testing provided insights into systematic understanding of parental involvement scenarios through SET (RQ1), parents’ perceptions of AI-generated content (RQ2), parents’ preferences for PAiREd during parent-AI collaboration in content creation (RQ3), and parent-robot collaboration in facilitating learning activities (RQ4). The insights reveal novel interaction paradigms for human-robot collaboration in the context of parenting and education. (3) Design Recommendations : We synthesized our findings into actionable recommendations to enhance parental involvement through AI-assisted learning companion robots. 2. Related Work In this section, we review research related to the importance and barriers to parental involvement; parental use of learning technologies; and the use of generative AI and robot in educational and parenting scenarios. 2.1. Importance and

Read the complete text