I don’t Want You to Die: A Shared Responsibility Framework for Safeguarding Child-Robot Companionship CCS: Human-centered computing Empirical studies in HCI CCS: Computer systems organization-Embedded and cyber-physical systems-Robotics
Fan Yang, Renkai Ma, Yaxin Hu, Michael D. Rodgers, Lingyao Li · 2025
Read I don’t Want You to Die: A Shared Responsibility Framework for Safeguarding Child-Robot Companionship CCS: Human-centered computing Empirical studies in HCI CCS: Computer systems organization-Embedded and cyber-physical systems-Robotics by Fan Yang, Renkai Ma, Yaxin Hu, Michael D. Rodgers, Lingyao Li (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-2510-26080.txt.
Introduction
Fan Yang, Renkai Ma, Yaxin Hu, Michael Rodgers, and Lingyao Li ask who should bear responsibility when a child robot companion is shut down. Social robots such as Moxie are built to form strong emotional bonds with children, so an abrupt end of service can cause real distress. The authors use the Moxie shutdown as a case study and survey seventy two United States participants in a qualitative study.
Yang and the team find that responsibility is seen as shared across the company, parents, developers, and government, though views split by political leaning and parental status. Opinions on whether service should continue are polarized, with supporters proposing technical, financial, or public paths and opponents citing business limits and risks of unhealthy dependence.
Catalog record
| CREATOR | Fan Yang, Renkai Ma, Yaxin Hu, Michael D. Rodgers, Lingyao Li |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 53 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-26080
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
I don’t Want You to Die: A Shared Responsibility Framework for Safeguarding Child-Robot Companionship CCS: Human-centered computing Empirical studies in HCI CCS: Computer systems organization-Embedded and cyber-physical systems-Robotics Fan Yang, Renkai Ma, Yaxin Hu, Michael D. Rodgers, Lingyao Li Abstract . Social robots like Moxie are designed to form strong emotional bonds with children, but their abrupt discontinuation can cause significant struggles and distress to children. When these services end, the resulting harm raises complex questions of who bears responsibility when children’s emotional bonds are broken. Using the Moxie shutdown as a case study through a qualitative survey of 72 U.S. participants, our findings show that the responsibility is viewed as a shared duty across the robot company, parents, developers, and government. However, these attributions varied by political ideology and parental status of whether they have children. Participants’ perceptions of whether the robot service should continue are highly polarized; supporters propose technical, financial, and governmental pathways for continuity, while opponents cite business realities and risks of unhealthy emotional dependency. Ultimately, this research contributes an empirically grounded shared responsibility framework for safeguarding child-robot companionship by detailing how accountability is distributed and contested, informing concrete design and policy implications to mitigate the emotional harm of robot discontinuation. 1. INTRODUCTION Humans can form strong emotional bonds with companion technologies, especially artificial intelligence (AI) and robotic companions, through emotional responsiveness and personalized interaction ( Adam 2025 ) . Users have been shown to develop deep emotional dependence and experience genuine grief when losing access to AI chatbots like Replika ( Laestadius et al. 2024 ; Ta et al. 2020 ; Banks 2024 ) . This bond is even more profound with embodied, physical robots, whose social presence fosters deeper, more sustained emotional connections through multi-sensory engagement and trust-building ( Lee et al. 2006 ; Li 2015 ; Wainer et al. 2006 ; Kiesler et al. 2008 ) . These companion robots are particularly appealing to vulnerable populations, especially children. For example, Research shows children, particularly those with neurological disorders, are highly likely to anthropomorphize their robot companions and form strong bonds with them ( Roštšinskaja et al. 2025 ; Holeva et al. 2024 ) , consistently describing them as best friends and family members ( Zhao and McEwen 2025 ; Constantinescu et al. 2022 ) . Children are drawn to social robots for the robots’ ability to engage and motivate interaction, their reduced social complexity relative to human agents, and the enhanced predictability of their behavioral responses ( Dubois-Sage et al. 2024 ; Hou et al. 2022 ) . Consequently, the market for social robots designed specifically for childhood companionship and developmental support has expanded rapidly in recent years ( Dawe et al. 2019 ) . Among these emerging products, the award-winning Moxie robot stands out as a prominent example ( Montalvo et al. 2024 ) . Developed by Embodied Inc., Moxie was designed and marketed to facilitate children’s acquisition of fundamental social-emotional competencies through play-based interactions, targeting skills such as turn-taking, eye contact, emotion regulation, and empathy ( Cano et al. 2021 ) . Standing at approximately one foot tall with a soft-touch blue body, Moxie was intentionally designed with large, expressive green eyes and a round, teardrop-shaped head that creates a non-threatening, approachable appearance. In addition, Moxie was also given an 8ish-year-old boy’s voice, making it especially easy to be identified by children as a friend rather than an authority figure (see Figure 1 ). Also, marketing strategies for such children’s companion robots uniformly target parental anxieties by promising “24/7” support with phrases like “always patient, always present” ( Gunawan et al. 2025 ) . These campaigns typically minimize the term “robot,” instead personifying the devices as constant “companions” or “friends” for children. Consequently, prominent examples like Moxie have engaged in over 4 million conversations since 2020, reaching tens of thousands of U.S. families ( Embodied, Inc. 2024 ) . (a) A child interacting with the Moxie robot on a desk, highlighting its role as an educational companion. (b) Various views of the Moxie robot with its design and features from different angles. (c) A child engaging with the Moxie robot in diverse home environments. Figure 1 . The Moxie companion robot in different interaction scenarios and showcasing its design. Unsurprisingly, when companion technologies fail, the emotional fallout is devastating ( Carter et al. 2020 ) . This phenomenon has been observed in adults grieving the loss of AI companions like Replika ( De Freitas et al. 2024 ; Namvarpour and Razi 2024 ) , but the emotional trauma is profoundly more severe for children, leaving parents at a loss in dealing with their grief ( Gunawan et al. 2025 ) . When Embodied Inc. abruptly shut down its Moxie robot, social media videos of children sobbing and pleading to “save” their friend vividly illustrated this distress. This burgeoning companion-robot industry, however, operates in an alarming regulatory vacuum, with its products falling through the cracks of laws that have not kept pace with technology ( De Freitas and Cohen 2025 ) . Because no single entity currently bears clear responsibility, the complexity of the Moxie robot ecosystem, involving manufacturers, developers, parents, and regulators, demands a comprehensive examination of a shared-responsibility framework to safeguard children’s emotional well-being ( McLennan and Eburn 2014 ) . Building on existing literature on shared responsibility in healthcare ( Elwyn et al. 2012 ; Bruch et al. 2024 ) and AI ethics ( Arroyo et al. 2022 ; Mühlhoff 2025 ; Lenk 2017 ) , our study employs the shutdown of the Moxie robot as a critical case to understand how such responsibility is attributed. To this end, we aim to investigate not only who the public holds accountable but also how these attributions differ across key demographics like parental status and political affiliation. Understanding these varied perspectives is important for developing an effective shared-responsibility framework and informing what further repair or coping is necessary for children and their families. Therefore, we ask two primary questions: RQ1: How is responsibility attributed for a child’s suffering after losing the Moxie companion robot? RQ1a: Who are the key stakeholders identified as responsible, and what are their assigned responsibilities? RQ1b: How might these attributions vary across demographic variables, like parental status and political affiliation? RQ2: How do people perceive the continuous service for the Moxie companion robot? RQ2a: Who are the key stakeholders responsible for this continuous service, and what are their responsibilities? RQ2b: Why do people oppose or express uncertainty about a continuous service? To answer these questions, we conducted a qualitative survey on N=72 U.S. participants. Through inductive thematic analysis on open-ended responses, we found that responsibility for a child’s suffering was attributed to multiple stakeholders, including the robot company, parents, developers, marketers, and the government (RQ1a). These attributions and proposed solutions demonstrated divergence based on political affiliation (systemic corporate failure vs. individual parental judgment) and parental status (hands-on parenting vs. foundational critique) (RQ1b). Besides, perceptions of continuous service for the Moxie robot were highly polarized, with support nearly matching opposition (RQ2b), leading participants to propose detailed pathways for continuity, including technical mechanisms, sustainable financial models, and external governmental guarantees (RQ2a). These findings inform our discussion of a shared