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

May combine two claims.

The study 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.

“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.”

May combine two claims.

Interviews about AI-generated content showed that parents value adaptability, efficiency, and affordability, but worry about age appropriateness, inaccuracies, and over-reliance.

“Second ( RQ2 ), interviews about AI-generated content showed that parents value adaptability, efficiency, and affordability, but worry about age appropriateness, inaccuracies, and over-reliance.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Most participating parents were mothers (17 mothers, 3 fathers).

“Most participating parents were mothers (17 mothers, 3 fathers), reflecting the greater likelihood of mothers serving as primary caregivers, consistent with previous research [ e.g., 66].”

May combine two claims.

The sociodemographic representation is limited: 85% of mothers held at least a Master's degree, and 80% of families reported an annual income of $100,000 or more.

“Despite efforts to recruit a diverse sample, the sociodemographic representation is limited: 85% of mothers held at least a Master's degree, and 80% of families reported an annual income of $100,000 or more.”

May combine two claims.

Integrating structured visual context significantly improved the accuracy of ChatGPT-4o to create question-answer pairs based on storybooks, reducing object-identification errors and enhancing consistency across pages.

“Integrating this structured approach significantly improved the accuracy of ChatGPT-4o to create question-answer pairs based on storybooks, reducing object-identification errors and enhancing consistency across pages.”

• AssumptionRests on an assumption: The ChatGPT-4o model faces limitations in quantitative and spatial reasoning on images and recognition of animated or anthropomorphic drawings, limiting its ability to accurately auto-generate question-answer pairs for literacy and math based on storybook images.

May combine two claims.

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).

“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).”

May combine two claims.

Parents co-created learning content for two books using PAiREd, indicating that difficulty, variety, and quality drive their evaluation of LLM-generated content.

“Third ( RQ3 ), parents co-created learning content for two books using PAiREd , indicating that difficulty, variety, and quality drive their evaluation of LLMgenerated content.”

May combine two claims.

Parents read the books with their children, adjusted role delegation between themselves and the robot according to the scenarios, and provided feedback.

“Finally ( RQ4 ), parents read the books with their children, adjusted role delegation between themselves and the robot according to the scenarios, and provided feedback.”

May combine two claims.

Many parents (16/20) were motivated when their child needed support, expressed interest and invited the parent to participate, or is well behaved and ready to learn.

“many parents (16/20) were motivated when their child needed support (P12, P15), expressed interest and invited the parent to participate (P1, P4, P7, P8, P10, P12-14, P16-19), or is well behaved and ready to learn (P2, P5, P13, P18, P19, P20).”

May combine two claims.

Others (5/20) lacked confidence in encouraging participation in learning activities or maintaining patience during learning support.

“Others (5/20) lacked confidence in encouraging participation in learning activities (P5, P11) or maintaining patience during learning support (P4, P10, P18, P19).”

May combine two claims.

When parents lack both energy and time, some chose robot takeover mode, adjusting their usage based on their trust in LLM-generated content, while others avoided using the system entirely due to low trust and insufficient time to review, or because the system design did not support independent use by young children.

“Some parents chose robot takeover mode , adjusting their usage based on their trust in LLMgenerated content. Others avoided using the system entirely due to low trust and insufficient time to review (P2, P12), or because the system design did not support independent use by young children (P4, P12, P20).”

May combine two claims.

When parents have sufficient energy but lack time, many opted for the robot takeover mode while adjusting their usage based on how much they trust LLM.

“many parents (P2, P3, P5, P7-11, P15) opted for the robot takeover mode -where the robot facilitates everything-while adjusting their usage based on how much they trust LLM .”

May combine two claims.

Parents who hold an attitude of open-minded caution recognize the risks of using AI-generated content but feel open to use it under specific conditions.

“On the other hand, parents who hold an attitude of open-minded caution recognize the risks of using AIgenerated content but feel open to use it under specific conditions.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A few parents noted AI's potential to foster affordability.

“a few parents (P1, P11, P18, P20) noted AI's potential to foster affordability , suggesting that AI-generated content could enhance the scalability and accessibility of learning resources, making ' more learning materials available, more variety available (P1), ' and making things ' cheaper and more accessible for people (P20).'”

Probably not a substantial claim: may not pass S3 (load-bearing).

A few parents expressed that social and public validation could enhance their trust in AI-generated content.

“a few parents (P2, P17, P19) expressed that social and public validation could also enhance their trust in AI-generated content, e.g., P2 described that ' if a thousand people used it...and endorsed this model, that would give me more confidence in it . '”

May combine two claims.

Some parents still prefer to skim through the content with minor editing even when time allows, as they found the detailed process too effortful but cannot fully trust LLM or themselves to come up with good questions.

“some parents (P1, P2, P7, P11, P12, P18) still prefer to skim through the content with minor editing , as they found the detailed process too effortful even when time allowed, but they cannot fully trust LLM or themselves to come up with good questions.”

May combine two claims.

Parents discussed customization as a benefit, creating material related to child's interests and motivating.

“parents (P2, P3, P4, P16, P19) discussed customization , illustrating that ' one of the big benefits would be to create material that are related to his[child's] interests and things that would be motivating to him[child] (P4). '”

May combine two claims.

Many parents (11/20) expressed concerns over repetitive content and preferred diverse topics to challenge their child differently.

“Many parents (11/20) expressed concerns over repetitive content and preferred diverse topics to challenge their child differently.”

May combine two claims.

Some parents involved their child in chores, but not all chores were seen as appropriate or safe for children.

“Although some parents involved their child in chores (P1, P2, P7, P10, P15, P17), not all chores were seen as being appropriate or safe for children.”

May combine two claims.

Many parents (8/20) avoided overly easy questions to prevent boredom but strategically included them at the start or after difficult questions to build confidence.

“Many (8/20) avoided overly easy questions to prevent boredom but strategically included them at the start or after difficult questions to build confidence.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A few parents highlighted the importance of expert involvement in creating AI-generated content.

“a few parents (P1, P15, P18) highlighted the importance of expert involvement in creating AI-generated content. For instance, P1 emphasized the need for oversight by ' people with a background in human development. '”

May combine two claims.

Some parents emphasized that the LLM output must be high-quality enough to require minimal editing, otherwise they may not use it at all.

“some parents (P4, P11, P15) emphasized that the LLM output must be high-quality enough to require minimal editing , otherwise they may not use it at all.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A few parents expected easier pedagogical integration with AI.

“a few parents (P14, P15) expected easier pedagogical integration with AI, enabling parents to ' teach children things that sometimes parents don't know because not all parents know everything (P14).'”

May combine two claims.

Parents' motivation to engage in learning activities is influenced by their physical and emotional state and their children's willingness to learn.

“we discovered that parents' motivation to engage in learning activities is influenced by their physical and emotional state and their children's willingness to learn”

May combine two claims.

Many parents expressed an interest in leveraging AI-generated content but also voiced concerns.

“In addition, many parents expressed an interest in leveraging content generated by AI but also voiced concerns, consistent with prior studies on parental control over media content [33, 80] and preferences for creating their own learning materials using AI [28].”

• cites [33] [33] Hui-Ru Ho, Edward M Hubbard, and Bilge Mutlu. 2024. " It's Not a Replacement:" Enabling Parent-Robot Collaboration to Support In-Home Learning Experiences of Young Children. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-18. - sourceNot checked yet

  Not checked in the free run.

• cites [80] [80] Junnan Yu, Xiang Qi, and Siqi Yang. 2024. Parent-Child Joint Media Engagement Within HCI: A Scoping Analysis of the Research Landscape. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-21. - sourceNot checked yet

  Not checked in the free run.

• cites [28] [28] Ariel Han, Xiaofei Zhou, Zhenyao Cai, Shenshen Han, Richard Ko, Seth Corrigan, and Kylie A Peppler. 2024. Teachers, Parents, and Students' perspectives on Integrating Generative AI into Elementary Literacy Education. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-17. -Not checked yet

  Not checked in the free run.

May combine two claims.

Many parents thought PAiREd has value in parenting education, providing them with pedagogical strategies and giving them opportunities to observe their child's proficiency level systematically through observation.

“Supported by mixed-method data, many parents thought PAiREd has value in parenting education, providing them with pedagogical strategies and giving them opportunities to observe their child's proficiency level systematically through observation.”

May combine two claims.

A few parents chose to avoid using the system entirely due to lack of trust in using LLM-generated content directly and insufficient time to review it, or because the system design did not support independent use for young children.

“a few parents (P2, P12) chose to avoid using the system entirely due to their lack of trust in using LLM-generated content directly and insufficient time to review it, or because the system design did not support independent use for young children (P4, P12, P20).”

May combine two claims.

Parents with less trust preferred to supervise while multitasking, review content beforehand using the editor, or use the system only if the LLM model met high-quality standards.

“parents with less trust preferred to supervise while multitasking (P2, P7, P8), review content beforehand using the editor (P2, P8), or use the system only if the LLM model met high-quality standards (P11, P15).”

May combine two claims.

Parents focus on balancing difficulty and variety of concepts as well as ensuring the quality of questions when reviewing, regenerating, and revising LLM-generated content for young children.

“Parents focus on balancing difficulty and variety of concepts as well as ensuring the quality of questions when reviewing, regenerating, and revising LLM-generated content for young children.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Parents described methods they envisioned to address their concerns.

“Parents described what methods they envisioned to address their concerns.”

Classified as probably not a substantial claim.

We discussed parents' preferences when collaborating with the AI-assisted robot across four contexts: sufficient energy and time, sufficient time but low energy, sufficient energy but limited time, and low energy and time.

“We discussed parents' preferences when collaborating with the AI-assisted robot across four contexts: (1) sufficient energy and time, (2) sufficient time but low energy, (3) sufficient energy but limited time, and (4) low energy and time.”

May combine two claims.

Parents perceive the value of the system to include content supervision, content co-creation with LLM, and parent empowerment through pedagogical insights.

“We found that parents perceive the value of the system to include not only content supervision , but also content co-creation with LLM and parent empowerment through pedagogical insights .”

May combine two claims.

When parents have sufficient time but lack energy, some opted for collaboration modes, with their involvement influenced by their motivation levels and partially by their skill relative to the robot.

“some parents (P3, P7, P10-12, P15, P18, P20) opted for collaboration modes , with their involvement influenced by their motivation levels and partially by their skill relative to the robot.”

The quote only partly states this claim.

A few parents preferred to avoid using the system entirely when time or energy was limited.

“A few parents (P1, P7) preferred to avoid using the system entirely , as they feel uncomfortable leaving their child engaged with the content without supervision.”

Classified as probably not a substantial claim.

We discussed parents' preferences and behaviors when collaborating with LLM under two main contexts: when parents have limited time or energy and when they have sufficient time and energy.

“We discussed parents' preferences and behaviors when collaborating with LLM under two main contexts: (1) when parents have limited time or energy and (2) when they have sufficient time and energy.”

May combine two claims.

Most parents (19/20) were not available when they needed to be at work or had other personal or professional engagements.

“Most parents (19/20) were not available when they needed to be at work (P1-4, P6, P7, P10-14, P16-19) and had other personal or professional engagements (P4, P5, P9, P11, P15, p20).”

May combine two claims.

When parents selected robot takeover mode, they wanted to either reduce their involvement while maintaining supervision or be completely absent.

“Third, when parents selected the robot takeover mode , they wanted to either reduce their involvement while maintaining supervision or be completely absent.”

May combine two claims.

Parents viewed the system not only as a tool for engaging their children but also for improving their own parenting skills (e.g., effective questioning technique and pedagogical strategies).

“Parents view the system not only as a tool for engaging their children, but also for improving their own parenting skills, e.g., effective questioning technique and pedagogical strategies.”

May combine two claims.

In the activity interface, various usage patterns were observed: parent takeover mode leveraging LLM content as support; collaborative modes determined by parent skill and motivation; robot takeover mode under supervision, with previewed content, or with original LLM content; and opting to avoid the system entirely.

“Similarly, in the activity interface, we observed various usage patterns: parent takeover mode leveraging LLM content as support; collaborative modes determined by parent skill and motivation; robot takeover mode under supervision, with previewed content, or with original LLM content; and opting to avoid the system entirely.”

Classified as probably not a substantial claim.

Many parents noted that older children might handle the system independently.

“Many parents noted that older children might handle the system independently, suggesting that the robot takeover mode should incorporate more interaction modalities, adjustable information density, and varying activity lengths to accommodate development needs of children as they mature.”

Classified as probably not a substantial claim.

P4 found parent takeover mode challenging because her child was too captivated and distracted by the robot, reducing her ability to engage fully.

“P4, for example, found this mode challenging because her child was too captivated and distracted by the robot, reducing her ability to engage fully.”

May combine two claims.

Even when parents had more time and energy, their approaches diverged: some enthusiastically customized AI-generated content, while others viewed customization as a burden and minimized or avoided using the system.

“Even when parents had more time and energy, their approaches diverged-some enthusiastically customized the AI-generated content, while others viewed customization as a burden and minimized or avoided using the system.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Parents, while familiar with preschool-level concepts, may struggle to generate questions spontaneously or accurately gauge their child's abilities.

“While familiar with preschool-level concepts, they may struggle to generate questions spontaneously or accurately gauge their child's abilities.”

May combine two claims.

In robot takeover mode, some parents preferred the robot to lead the activity while they remained nearby to assist if needed; others wanted the robot to manage the entire activity independently for the child.

“In the first case, they preferred the robot to lead the activity while they remained nearby to assist if needed. In the second, they wanted the robot to manage the entire activity independently for the child.”

Probably not a substantial claim: may not pass S4 (based).

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.

“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.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

AI-driven systems can support parenting education and provide evidence-based advice through applications and chatbots, delivering micro-interventions.

“Prior work suggested that AI-driven systems can support parenting education [61] and provide evidence-based advice through applications and chatbots, delivering micro-interventions such as teaching parents how to praise their children effectively [11, 17] or offering strategies to teach complex concepts [55, 69].”

• cites [61] [61] Melina Petsolari, Seray B Ibrahim, and Petr Slovak. 2024. Socio-technical Imaginaries: Envisioning and Understanding AI Parenting Supports through Design Fiction. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-27. - sourceNot checked yet

  Not checked in the free run.

• cites [11] [11] Deborah Winders Davis, M Cynthia Logsdon, Krista Vogt, Jeff Rushton, John Myers, Adrian Lauf, and Felicia Hogan. 2017. Parent education is changing: a review of smartphone apps. MCN: The American Journal of Maternal/Child Nursing 42, 5 (2017), 248-256. - sourceNot checked yet

  Not checked in the free run.

• cites [17] [17] Guido A Entenberg, Gilly Dosovitsky, Shirin Aghakhani, Karin Mostovoy, N Carre, Zendrea Marshall, D Benfica, S Mizrahi, A Testerman, A Rousseau, et al. 2023. User experience with a parenting chatbot micro intervention. Frontiers in Digital Health 4 (2023), 989022. - sourceNot checked yet

  Not checked in the free run.

• cites [55] [55] Reza Hadi Mogavi, Chao Deng, Justin Juho Kim, Pengyuan Zhou, Young D Kwon, Ahmed Hosny Saleh Metwally, Ahmed Tlili, Simone Bassanelli, Antonio Bucchiarone, Sujit Gujar, et al. 2024. ChatGPT in education: A blessing or a curse? A qualitative study exploring early adopters' utilization and perceptions. Computers in Human Behavior: Artificial Humans 2, 1 (2024), 100027. - sourceNot checked yet

  Not checked in the free run.

• cites [69] [69] Jiahong Su and Weipeng Yang. 2023. Unlocking the power of ChatGPT: A framework for applying generative AI in education. ECNU Review of Education 6, 3 (2023), 355-366. - sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

Parental supervision is needed to ensure alignment of AI and robots with parents' educational goals.

“emphasizing the need for parental supervision to ensure alignment of AI and robots with parents' educational goals [28, 33].”

• cites [28] [28] Ariel Han, Xiaofei Zhou, Zhenyao Cai, Shenshen Han, Richard Ko, Seth Corrigan, and Kylie A Peppler. 2024. Teachers, Parents, and Students' perspectives on Integrating Generative AI into Elementary Literacy Education. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-17. -Not checked yet

  Not checked in the free run.

• cites [33] [33] Hui-Ru Ho, Edward M Hubbard, and Bilge Mutlu. 2024. " It's Not a Replacement:" Enabling Parent-Robot Collaboration to Support In-Home Learning Experiences of Young Children. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-18. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

According to parental mediation theory, parents mitigate risks through restrictive mediation, active mediation, and co-use.

“According to parental mediation theory, parents mitigate these risks through restrictive mediation (e.g., setting limits), active mediation (e.g., discussing content), and co-use (e.g., shared use of technology) [74].”

• cites [74] [74] Patti M Valkenburg, Marina Krcmar, Allerd L Peeters, and Nies M Marseille. 1999. Developing a scale to assess three styles of television mediation:'Instructive mediation,''restrictive mediation,' and 'social coviewing'. Journal of broadcasting & electronic media 43, 1 (1999), 52-66. - sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim. May combine two claims.

Technology encourages parental involvement by facilitating parent-child engagement in learning activities while introducing risks that require active parental mediation.

“Technology encourages parental involvement by facilitating parentchild engagement in learning activities while introducing risks that require active parental mediation [21].”

• cites [21] [21] Alyssa R Gonzalez-DeHass, Patricia P Willems, Jillian R Powers, and Ann T Musgrove. 2022. Parental involvement in supporting students' digital learning. Educational Psychologist 57, 4 (2022), 281-294. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

AI-powered systems can personalize educational experiences to address individual learning needs.

“Prior work has highlighted that AI-powered systems can personalize educational experiences to address individual learning needs [78, 79, 82]”

• cites [78] [78] Ying Xu, Kunlei He, Julian Levine, Daniel Ritchie, Zexuan Pan, Andres Bustamante, and Mark Warschauer. 2024. Artificial intelligence enhances children's science learning from television shows. Journal of Educational Psychology (2024). -Not checked yet

  Not checked in the free run.

• cites [79] [79] Ying Xu, Valery Vigil, Andres S Bustamante, and Mark Warschauer. 2022. 'Elinor's Talking to Me!': Integrating Conversational AI into Children's Narrative Science Programming. In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems . 1-16. - sourceNot checked yet

  Not checked in the free run.

• cites [82] [82] Chao Zhang, Xuechen Liu, Katherine Ziska, Soobin Jeon, Chi-Lin Yu, and Ying Xu. 2024. Mathemyths: leveraging large language models to teach mathematical language through Child-AI co-creative storytelling. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-23. - sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Technology introduces risks such as excessive screen time, exposure to inappropriate content, and misinformation, which necessitate parental intervention.

“On the negative side, technology introduces risks such as excessive screen time, exposure to inappropriate content, and misinformation, which necessitate parental intervention [37, 58].”

• cites [37] [37] Philip N Howard, Lisa-Maria Neudert, Nayana Prakash, and Steven Vosloo. 2021. Digital misinformation/disinformation and children. UNICEF. Retrieved on February 20 (2021), 2021. -Not checked yet

  Not checked in the free run.

• cites [58] [58] Tassia K Oswald, Alice R Rumbold, Sophie GE Kedzior, and Vivienne M Moore. 2020. Psychological impacts of 'screen time' and 'green time' for children and adolescents: A systematic scoping review. PloS one 15, 9 (2020), e0237725. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

GAI offers promising support for parents by enhancing their ability to educate and engage with their children.

“GAI offers promising support for parents by enhancing their ability to educate and engage with their children.”

• cites [61] [61] Melina Petsolari, Seray B Ibrahim, and Petr Slovak. 2024. Socio-technical Imaginaries: Envisioning and Understanding AI Parenting Supports through Design Fiction. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-27. - sourceNot checked yet

  Not checked in the free run.

• cites [11] [11] Deborah Winders Davis, M Cynthia Logsdon, Krista Vogt, Jeff Rushton, John Myers, Adrian Lauf, and Felicia Hogan. 2017. Parent education is changing: a review of smartphone apps. MCN: The American Journal of Maternal/Child Nursing 42, 5 (2017), 248-256. - sourceNot checked yet

  Not checked in the free run.

• cites [17] [17] Guido A Entenberg, Gilly Dosovitsky, Shirin Aghakhani, Karin Mostovoy, N Carre, Zendrea Marshall, D Benfica, S Mizrahi, A Testerman, A Rousseau, et al. 2023. User experience with a parenting chatbot micro intervention. Frontiers in Digital Health 4 (2023), 989022. - sourceNot checked yet

  Not checked in the free run.

• cites [55] [55] Reza Hadi Mogavi, Chao Deng, Justin Juho Kim, Pengyuan Zhou, Young D Kwon, Ahmed Hosny Saleh Metwally, Ahmed Tlili, Simone Bassanelli, Antonio Bucchiarone, Sujit Gujar, et al. 2024. ChatGPT in education: A blessing or a curse? A qualitative study exploring early adopters' utilization and perceptions. Computers in Human Behavior: Artificial Humans 2, 1 (2024), 100027. - sourceNot checked yet

  Not checked in the free run.

• cites [69] [69] Jiahong Su and Weipeng Yang. 2023. Unlocking the power of ChatGPT: A framework for applying generative AI in education. ECNU Review of Education 6, 3 (2023), 355-366. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Having access to computers at home significantly boosts academic achievement of young children when parents actively mediate their use.

“In addition, having access to computers at home significantly boost academic achievement of young children when parents actively mediate their use [18, 35].”

• cites [18] [18] Linda M Espinosa, James M Laffey, Tiffany Whittaker, and Yanyan Sheng. 2006. Technology in the home and the achievement of young children: Findings from the early childhood longitudinal study. Early education and development 17, 3 (2006), 421-441. - sourceNot checked yet

  Not checked in the free run.

• cites [35] [35] Sandra L Hofferth. 2010. Home media and children's achievement and behavior. Child development 81, 5 (2010), 1598-1619. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The CAM framework suggests parents delegate tasks to a robot when they feel less capable, have limited time, or are unmotivated.

“The CAM framework similarly suggests parents delegate tasks to a robot when they feel less capable, have limited time, or are unmotivated.”

• cites [33] [33] Hui-Ru Ho, Edward M Hubbard, and Bilge Mutlu. 2024. " It's Not a Replacement:" Enabling Parent-Robot Collaboration to Support In-Home Learning Experiences of Young Children. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-18. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

e-books promote interactive behaviors between parents and children better than print books.

“For example, e-books promote interactive behaviors between parents and children better than print books [43].”

• cites [43] [43] Ofra Korat and Tal Or. 2010. How new technology influences parent-child interaction: The case of e-book reading. First Language 30, 2 (2010), 139-154. - sourceNot checked yet

  Not checked in the free run.

May combine two claims.

The limited sociodemographic representation limits the generalizability of the findings and may omit insights into underrepresented groups.

“This limits the generalizability of our findings and may omit insights into underrepresented groups, a limitation further discussed in Section 8.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The greater likelihood of mothers serving as primary caregivers is consistent with previous research.

“Most participating parents were mothers (17 mothers, 3 fathers), reflecting the greater likelihood of mothers serving as primary caregivers, consistent with previous research [ e.g., 66].”

• cites [ e.g., 66]Not checked yet

  Not checked in the free run.

May combine two claims.

Unlike teachers, parents often lack formal pedagogical training and may underestimate their role in supporting children's learning, particularly as young children struggle to articulate their needs.

“Unlike teachers, parents often lack formal pedagogical training and may underestimate their role in supporting children's learning, particularly as young children struggle to articulate their needs [29].”

• cites [29] [29] Steven R Hara and Daniel J Burke. 1998. Parent involvement: The key to improved student achievement. School community journal 8, 2 (1998), 9-19. -Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Mobile applications can improve learning outcomes but require parents to possess sufficient technology efficacy to guide their use.

“Mobile applications, for instance, can improve learning outcomes but require parents to possess sufficient technology efficacy to guide their use [59].”

• cites [59] [59] Stamatios Papadakis, Nikolas Zaranis, and Michail Kalogiannakis. 2019. Parental involvement and attitudes towards young Greek children's mobile usage. International Journal of Child-Computer Interaction 22 (2019), 100144. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Many parents prefer using GAI to create educational materials tailored to their children's needs, rather than granting children direct access to these tools.

“Many parents also prefer using GAI to create educational materials tailored to their children's needs, rather than granting children direct access to these tools [27].”

• cites [27] [27] Ariel Han and Zhenyao Cai. 2023. Design implications of generative AI systems for visual storytelling for young learners. In Proceedings of the 22nd Annual ACM Interaction Design and Children Conference . 470-474. - sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

The qualitative data analysis used reflexive Thematic Analysis following Clarke and Braun [9] and McDonald et al. [52].

“For qualitative data, we conducted a reflexive Thematic Analysis (TA) following Clarke and Braun [9] and McDonald et al. [52] to identify patterns and understand parental perspectives.”

• cites [9] [9] Victoria Clarke and Virginia Braun. 2014. Thematic analysis. In Encyclopedia of critical psychology . Springer, 1947-1952. - sourceNot checked yet

  Not checked in the free run.

• cites [52] [52] Nora McDonald, Sarita Schoenebeck, and Andrea Forte. 2019. Reliability and inter-rater reliability in qualitative research: Norms and guidelines for CSCW and HCI practice. Proceedings of the ACM on human-computer interaction 3, CSCW (2019), 1-23. - sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Many parents are skeptical about the role of AI in their children's academic processes, concerned about the accuracy of AI-generated content, and worry that over-reliance on AI could stifle independent thinking.

“Many parents are skeptical about the role of AI in their children's academic processes, concerned about the accuracy of AI-generated content, and worry that over-reliance on AI could stifle independent thinking [27].”

• cites [27] [27] Ariel Han and Zhenyao Cai. 2023. Design implications of generative AI systems for visual storytelling for young learners. In Proceedings of the 22nd Annual ACM Interaction Design and Children Conference . 470-474. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Companion robots can effectively support both school-based learning (e.g., math, literacy, science) and home-based learning activities (e.g., reading, number board games, math-oriented conversations with parents).

“Prior work demonstrated that companion robots can effectively support both school-based learning ( e.g., math [48], literacy [22, 41], and science [12]) and home-based learning activities ( e.g., reading [53, 54], number board games [32], and math-oriented conversations with parents [34]).”

• cites [48] [48] Edgar Lopez-Caudana, Pedro Ponce, Luis Cervera, Sara Iza, and Nancy Mazon. 2018. Robotic platform for teaching maths in junior high school. International Journal on Interactive Design and Manufacturing (IJIDeM) 12, 4 (2018), 1349-1360. - sourceNot checked yet

  Not checked in the free run.

• cites [22] [22] Goren Gordon, Samuel Spaulding, Jacqueline Kory Westlund, Jin Joo Lee, Luke Plummer, Marayna Martinez, Madhurima Das, and Cynthia Breazeal. 2016. Affective personalization of a social robot tutor for children's second language skills. In Proceedings of the AAAI conference on artificial intelligence , Vol. 30. -Not checked yet

  Not checked in the free run.

• cites [41] [41] James Kennedy, Paul Baxter, Emmanuel Senft, and Tony Belpaeme. 2016. Social robot tutoring for child second language learning. In 2016 11th ACM/IEEE international conference on human-robot interaction (HRI) . IEEE, 231-238. -Not checked yet

  Not checked in the free run.

• cites [12] [12] Daniel P Davison, Frances M Wijnen, Vicky Charisi, Jan van der Meij, Vanessa Evers, and Dennis Reidsma. 2020. Working with a social robot in school: a longterm real-world unsupervised deployment. In Proceedings of the 2020 ACM/IEEE International Conference on Human-Robot Interaction . 63-72. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Parents envisioned robots as their collaborators to support their children's learning at home, and their collaboration patterns can be determined by the parents' capability, availability, and motivation.

“Beyond educational support, Ho et al. [33] uncovered that parents envisioned robots as their collaborators to support their children's learning at home and that their collaboration patterns can be determined by the parents' capability, availability, and motivation.”

• cites [33] [33] Hui-Ru Ho, Edward M Hubbard, and Bilge Mutlu. 2024. " It's Not a Replacement:" Enabling Parent-Robot Collaboration to Support In-Home Learning Experiences of Young Children. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-18. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

In parent-toddler-robot interaction, parents tend to decrease their scaffolding affectively when the robot increases its scaffolding behavior.

“Gvirsman and Gordon [25] found that, in a parent-toddler-robot interaction, parents tend to decrease their scaffolding affectively when the robot increases its scaffolding behavior.”

• cites [25] [25] Omer Gvirsman and Goren Gordon. 2024. Effect of Social Robot's Role and Behavior on Parent-Toddler Interaction. In Proceedings of the 2024 ACM/IEEE - International Conference on Human-Robot Interaction . 222-230. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

A LLM-driven conversational system can teach children mathematical concepts through co-creative storytelling, achieving learning outcomes similar to human-led instructions.

“For example, a LLM-driven conversational system can teach children mathematical concepts through co-creative storytelling, achieving learning outcomes similar to human-led instructions [82].”

• cites [82] [82] Chao Zhang, Xuechen Liu, Katherine Ziska, Soobin Jeon, Chi-Lin Yu, and Ying Xu. 2024. Mathemyths: leveraging large language models to teach mathematical language through Child-AI co-creative storytelling. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-23. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The first three authors were experienced in qualitative coding.

“The first three authors, experienced in qualitative coding, transcribed and familiarized themselves with the data from audio and video recordings.”

Classified as probably not a substantial claim. May combine two claims.

The remaining data were independently coded by the three authors, with peer reviews and iterative discussions to finalize codes.

“Using the codebook, the remaining data were independently coded by the three authors, with peer reviews and iterative discussions to finalize codes.”

May combine two claims.

Zhang et al. emphasized the importance of flexible parental involvement during reading through Storybuddy, but focused on a virtual chatbot rather than a physical robot, and how flexible modes may be used in different scenarios remain unknown.

“Though Zhang et al. [83] emphasized the importance of flexible parental involvement during reading through a system called Storybuddy , yet they focused on a virtual chatbot rather than a physical robot, and how the flexible modes may be used in different scenarios remain unknown.”

• cites [83] [83] Zheng Zhang, Ying Xu, Yanhao Wang, Bingsheng Yao, Daniel Ritchie, Tongshuang Wu, Mo Yu, Dakuo Wang, and Toby Jia-Jun Li. 2022. Storybuddy: A human-ai collaborative chatbot for parent-child interactive storytelling with flexible parental involvement. In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems . 1-21. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Social robots enhanced parent-child co-reading activities.

“Similarly, Chen et al. [7] found that social robots enhanced parent-child co-reading activities”

• cites [7] [7] Huili Chen, Anastasia K Ostrowski, Soo Jung Jang, Cynthia Breazeal, and Hae Won Park. 2022. Designing Long-term Parent-child-robot Triadic Interaction at Home through Lived Technology Experiences and Interviews. In 2022 31st IEEE International Conference on Robot and Human Interactive Communication (RO-MAN) . IEEE, 401-408. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The robotic system Patricc fostered more triadic interaction between parents and toddlers than a tablet.

“Gvirsman et al. [26] showed that the robotic system, Patricc , fostered more triadic interaction between parents and toddlers than a tablet”

• cites [26] [26] Omer Gvirsman, Yaacov Koren, Tal Norman, and Goren Gordon. 2020. Patricc: A platform for triadic interaction with changeable characters. In Proceedings of the 2020 ACM/IEEE International Conference on Human-robot Interaction . 399-407. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

There is a significant risk of GAI producing inaccurate or biased information, discouraging independent thought among children, and threatening user privacy.

“There is also a significant risk of GAI producing inaccurate or biased information, discouraging independent thought among children, and threatening user privacy [27, 28, 69].”

• cites [27] [27] Ariel Han and Zhenyao Cai. 2023. Design implications of generative AI systems for visual storytelling for young learners. In Proceedings of the 22nd Annual ACM Interaction Design and Children Conference . 470-474. - sourceNot checked yet

  Not checked in the free run.

• cites [28] [28] Ariel Han, Xiaofei Zhou, Zhenyao Cai, Shenshen Han, Richard Ko, Seth Corrigan, and Kylie A Peppler. 2024. Teachers, Parents, and Students' perspectives on Integrating Generative AI into Elementary Literacy Education. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-17. -Not checked yet

  Not checked in the free run.

• cites [69] [69] Jiahong Su and Weipeng Yang. 2023. Unlocking the power of ChatGPT: A framework for applying generative AI in education. ECNU Review of Education 6, 3 (2023), 355-366. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Parents generally have positive attitudes toward incorporating robots into their children's learning, but remain concerned about the risk of disrupting school-based learning and potential teacher replacement.

“Although parents generally have positive attitudes toward incorporating robots into their children's learning, they remain concerned about the risk of disrupting school-based learning and potential teacher replacement [45, 49, 73].”

• cites [45] [45] EunKyoung Lee, YoungJun Lee, Bokyung Kye, and Beomseog Ko. 2008. Elementary and middle school teachers', students' and parents' perception of robot-aided education in Korea. In EdMedia+ Innovate Learning . Association for the Advancement of Computing in Education (AACE), 175-183. -Not checked yet

  Not checked in the free run.

• cites [49] [49] Belinda Louie, Elin A Björling, and Annie Camey Kuo. 2021. The desire for social robots to support English language learners: Exploring robot perceptions of teachers, parents, and students. In Frontiers in Education , Vol. 6. Frontiers Media SA, 566909. - sourceNot checked yet

  Not checked in the free run.

• cites [73] [73] Nils F Tolksdorf and Katharina J Rohlfing. 2020. Parents' views on using social robots for language learning. In 2020 29th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN) . IEEE, 634-640. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The team collaboratively coded data from P1 using the initial codebook, discussed interpretations, and created a consensus-based codebook.

“The team collaboratively coded data from P1 using the initial codebook, discussed interpretations, and created a consensus-based codebook.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Physically embodied robots provide adaptive assistance and verbal interaction similar to virtual or conversational agents, yet they foster greater engagement with the physical environment and encourage more advanced social behaviors during learning, leading to improved learning outcomes.

“Physically embodied robots provide adaptive assistance and verbal interaction similar to virtual or conversational agents [5, 46, 63, 65], yet they foster greater engagement with the physical environment and encourage more advanced social behaviors during learning [2], leading to improved learning outcomes [47].”

• cites [5] [5] LaVonda N Brown and Ayanna M Howard. 2014. The positive effects of verbal encouragement in mathematics education using a social robot. In 2014 IEEE integrated STEM education conference . IEEE, 1-5. - sourceNot checked yet

  Not checked in the free run.

• cites [46] [46] Daniel Leyzberg, Samuel Spaulding, and Brian Scassellati. 2014. Personalizing robot tutors to individuals' learning differences. In Proceedings of the 2014 ACM/IEEE international conference on Human-robot interaction . 423-430. - sourceNot checked yet

  Not checked in the free run.

• cites [63] [63] Aditi Ramachandran, Sarah Strohkorb Sebo, and Brian Scassellati. 2019. Personalized robot tutoring using the assistive tutor pOMDP (AT-POMDP). In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 33. 8050-8057. - sourceNot checked yet

  Not checked in the free run.

• cites [65] [65] Thorsten Schodde, Kirsten Bergmann, and Stefan Kopp. 2017. Adaptive robot language tutoring based on Bayesian knowledge tracing and predictive decisionmaking. In Proceedings of the 2017 ACM/IEEE International Conference on HumanRobot Interaction . 128-136. - sourceNot checked yet

  Not checked in the free run.

• cites [2] [2] Tony Belpaeme, James Kennedy, Aditi Ramachandran, Brian Scassellati, and Fumihide Tanaka. 2018. Social robots for education: A review. Science robotics 3, 21 (2018), eaat5954. - sourceNot checked yet

  Not checked in the free run.

• cites [47] [47] Daniel Leyzberg, Samuel Spaulding, Mariya Toneva, and Brian Scassellati. 2012. The physical presence of a robot tutor increases cognitive learning gains. In Proceedings of the annual meeting of the cognitive science society , Vol. 34. -Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

GAI has limitations such as limited effectiveness in more complex learning tasks, limited quality of training data, and inability to offer comprehensive educational support.

“Prior work highlighted the limitations of GAI, such as its limited effectiveness in more complex learning tasks,the limited quality of the training data, and its inability to offer comprehensive educational support [69].”

• cites [69] [69] Jiahong Su and Weipeng Yang. 2023. Unlocking the power of ChatGPT: A framework for applying generative AI in education. ECNU Review of Education 6, 3 (2023), 355-366. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

AI-based storytelling tools address practical challenges such as time constraints by alleviating physical labor while fostering parent-child interactions.

“Beyond educational support, AI-based storytelling tools address practical challenges ( e.g., time constraints) by alleviating physical labor while fostering parent-child interactions [71].”

• cites [71] [71] Yuling Sun, Jiali Liu, Bingsheng Yao, Jiaju Chen, Dakuo Wang, Xiaojuan Ma, Yuxuan Lu, Ying Xu, and Liang He. 2024. Exploring Parent's Needs for ChildrenCentered AI to Support Preschoolers' Storytelling and Reading Activities. arXiv preprint arXiv:2401.13804 (2024). -Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Children can learn elements of a second language from a robot in short-term interactions.

“For example, Kennedy et al. [41] suggested that children can learn elements of a second language from a robot in short-term interactions”

• cites [41] [41] James Kennedy, Paul Baxter, Emmanuel Senft, and Tony Belpaeme. 2016. Social robot tutoring for child second language learning. In 2016 11th ACM/IEEE international conference on human-robot interaction (HRI) . IEEE, 231-238. -Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Children who took on the role of teaching the robot gained confidence and improved learning outcomes.

“and Tanaka and Kimura [72] found that children who took on the role of teaching the robot gained confidence and improved learning outcomes.”

• cites [72] [72] Fumihide Tanaka and Takeshi Kimura. 2009. The use of robots in early education: a scenario based on ethical consideration. In RO-MAN 2009-The 18th IEEE International Symposium on Robot and Human Interactive Communication . IEEE, 558-560. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

GAI can help create personalized learning experiences by providing timely feedback and tailoring content, enhancing positive learning experiences.

“It can help create personalized learning experiences by providing timely feedback and tailoring content [28, 55, 69], enhancing positive learning experiences [39].”

• cites [28] [28] Ariel Han, Xiaofei Zhou, Zhenyao Cai, Shenshen Han, Richard Ko, Seth Corrigan, and Kylie A Peppler. 2024. Teachers, Parents, and Students' perspectives on Integrating Generative AI into Elementary Literacy Education. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-17. -Not checked yet

  Not checked in the free run.

• cites [55] [55] Reza Hadi Mogavi, Chao Deng, Justin Juho Kim, Pengyuan Zhou, Young D Kwon, Ahmed Hosny Saleh Metwally, Ahmed Tlili, Simone Bassanelli, Antonio Bucchiarone, Sujit Gujar, et al. 2024. ChatGPT in education: A blessing or a curse? A qualitative study exploring early adopters' utilization and perceptions. Computers in Human Behavior: Artificial Humans 2, 1 (2024), 100027. - sourceNot checked yet

  Not checked in the free run.

• cites [69] [69] Jiahong Su and Weipeng Yang. 2023. Unlocking the power of ChatGPT: A framework for applying generative AI in education. ECNU Review of Education 6, 3 (2023), 355-366. - sourceNot checked yet

  Not checked in the free run.

• cites [39] [39] Jussi S Jauhiainen and Agustín Garagorry Guerra. 2023. Generative AI and ChatGPT in school children's education: Evidence from a school lesson. Sustainability 15, 18 (2023), 14025. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The WAKEY robot improved morning routines and reduced parental frustration.

“while Chan et al. [6] demonstrated that the WAKEY robot improved morning routines and reduced parental frustration.”

• cites [6] [6] Meng-Ying Chan, Yi-Hsuan Lin, Long-Fei Lin, Ting-Wei Lin, Wei-Che Hsu, Chiayu Chang, Rui Liu, Ko-Yu Chang, Min-hua Lin, and Jane Yung-jen Hsu. 2017. WAKEY: assisting parent-child communication for better morning routines. In Proceedings of the 2017 ACM Conference on Computer Supported Cooperative Work and Social Computing . 2287-2299. - sourceNot checked yet

  Not checked in the free run.

May combine two claims.

ContextQ presented auto-generated dialogic questions to caregivers for dialogic reading, but primarily considered situations where parents are actively involved, not scenarios where parents cannot participate fully.

“Similarly, ContextQ [13] presented auto-generated dialogic questions to caregivers for dialogic reading, but primarily considered situations where parents are actively involved, not scenarios where parents cannot participate fully.”

• cites [13] [13] Griffin Dietz Smith, Siddhartha Prasad, Matt J Davidson, Leah Findlater, and RBenjamin Shapiro. 2024. ContextQ: Generated Questions to Support Meaningful Parent-Child Dialogue While Co-Reading. In Proceedings of the 23rd Annual ACM Interaction Design and Children Conference . 408-423. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Prior research has outlined five key principles for age-appropriate AI: fairness and equality, transparency and accountability, privacy and exploitation, safety and safeguarding, and sustainability and age-appropriateness.

“To mitigate harm from AI for children, prior research has outlined five key principles for age-appropriate AI: fairness and equality, transparency and accountability, privacy and exploitation, safety and safeguarding, and sustainability and age-appropriateness [75].”

• cites [75] [75] Ge Wang, Jun Zhao, Max Van Kleek, and Nigel Shadbolt. 2022. Informing ageappropriate ai: Examining principles and practices of ai for children. In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems . 1-29. - sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

Preschoolers benefit from exposure to advanced content in literacy and math, compared to basic content.

“Claessens et al. [8] and Engel et al. [16] suggested that preschoolers benefit from exposure to advanced content in literacy and math, compared to basic content.”

• cites [8] [8] AmyClaessens, Mimi Engel, and F Chris Curran. 2014. Academic content, student learning, and the persistence of preschool effects. American Educational Research Journal 51, 2 (2014), 403-434. - sourceNot checked yet

  Not checked in the free run.

• cites [16] [16] Mimi Engel, Amy Claessens, and Maida A Finch. 2013. Teaching students what they already know? The (mis) alignment between mathematics instructional content and student knowledge in kindergarten. Educational Evaluation and Policy Analysis 35, 2 (2013), 157-178. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Parents have challenges in distinguishing elementary-school students' authorship and controlling misinformation, as noted by Han et al.

“Similar concerns were raised by Han et al. [28], who noted parents' challenges in distinguishing elementary-school students' authorship and controlling misinformation.”

• cites [28] [28] Ariel Han, Xiaofei Zhou, Zhenyao Cai, Shenshen Han, Richard Ko, Seth Corrigan, and Kylie A Peppler. 2024. Teachers, Parents, and Students' perspectives on Integrating Generative AI into Elementary Literacy Education. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-17. -Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

Supporting parents with pedagogical strategies in advanced content and providing them with opportunities to observe their children's developmental progress can further optimize the value of parental involvement.

“Therefore, supporting parents with pedagogical strategies in advanced content and providing them with opportunities to observe their children's developmental progress can further optimize the value of parental involvement.”

May combine two claims.

If the system provides a comprehensive framework and ample ideas, parents may not decrease their involvement in an activity just because they don't have the pedagogical skill; in fact, they may even increase their involvement.

“If the system provides a comprehensive framework and ample ideas, parents may not decrease their involvement in an activity just because they don't have the pedagogical skill; in fact, they may even increase their involvement.”

Probably not a substantial claim: may not pass S3 (load-bearing).

These results exemplify prior parental involvement frameworks.

“These results exemplify prior parental involvement frameworks [23, 33]”

• cites [23] [23] Christa L Green, Joan MT Walker, Kathleen V Hoover-Dempsey, and Howard M Sandler. 2007. Parents' motivations for involvement in children's education: An empirical test of a theoretical model of parental involvement. Journal of educational psychology 99, 3 (2007), 532. - sourceNot checked yet

  Not checked in the free run.

• cites [33] [33] Hui-Ru Ho, Edward M Hubbard, and Bilge Mutlu. 2024. " It's Not a Replacement:" Enabling Parent-Robot Collaboration to Support In-Home Learning Experiences of Young Children. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1-18. - sourceNot checked yet

  Not checked in the free run.

May combine two claims.

While parents generally have the skill to support their children's development, they seek more guidance on pedagogical strategies, especially for advanced concepts.

“We found that while parents generally have the skill to support their children's development, they seek more guidance on pedagogical strategies, especially for advanced concepts.”

Probably not a substantial claim: may not pass S4 (based).

The current design presents content uniformly across modes.

“Yet, the current design presents content uniformly across modes.”

May combine two claims.

A predominance of mothers were included due to self-selection bias, consistent with previous education studies, and the study faced difficulties recruiting lower-income families.

“Consistent with previous education studies, a predominance of mothers were included due to self-selection bias [51, 66], and we faced difficulties recruiting lower-income families [56].”

• cites [51] [51] Brent A McBride and Gail Mills. 1993. A comparison of mother and father involvement with their preschool age children. Early childhood research quarterly 8, 4 (1993), 457-477. - sourceNot checked yet

  Not checked in the free run.

• cites [66] [66] Sarah J Schoppe-Sullivan, Letitia E Kotila, Rongfang Jia, Sarah N Lang, and Daniel J Bower. 2013. Comparisons of levels and predictors of mothers' and fathers' engagement with their preschool-aged children. Early child development and care 183, 3-4 (2013), 498-514. - sourceNot checked yet

  Not checked in the free run.

• cites [56] [56] Lisa M Nicholson, Patricia M Schwirian, Elizabeth G Klein, Theresa Skybo, Lisa Murray-Johnson, Ihuoma Eneli, Bethany Boettner, Gina M French, and Judith A Groner. 2011. Recruitment and retention strategies in longitudinal clinical studies with low-income populations. Contemporary clinical trials 32, 3 (2011), 353-362. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable).

The current presentation of information may not optimally support parents in leading these activities.

“The current presentation of information may not optimally support parents in leading these activities.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The system currently targets literacy and math.

“Furthermore, our system currently targets literacy and math.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Some educational robots, such as Miko, are now priced similarly to smartphones, improving feasibility for families.

“In addition, while cost remains a common accessibility limitation, some educational robots, such as Miko, 13 are now priced similarly to smartphones, improving feasibility for families.”

• cites [13] [13] Griffin Dietz Smith, Siddhartha Prasad, Matt J Davidson, Leah Findlater, and RBenjamin Shapiro. 2024. ContextQ: Generated Questions to Support Meaningful Parent-Child Dialogue While Co-Reading. In Proceedings of the 23rd Annual ACM Interaction Design and Children Conference . 408-423. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Zhang et al. emphasized the importance of flexible parental involvement during reading through Storybuddy, and Dietz Smith et al. presented auto-generated dialogic questions to caregivers for dialogic reading similar to the paper's parent-takeover mode.

“Zhang et al. [83] similarly emphasized the importance of flexible parental involvement during reading through a system called Storybuddy using a virtual chatbot, while Dietz Smith et al. [13] presented auto-generated dialogic questions to caregivers for dialogic reading that is similar to our parent-takeover mode.”

• cites [83] [83] Zheng Zhang, Ying Xu, Yanhao Wang, Bingsheng Yao, Daniel Ritchie, Tongshuang Wu, Mo Yu, Dakuo Wang, and Toby Jia-Jun Li. 2022. Storybuddy: A human-ai collaborative chatbot for parent-child interactive storytelling with flexible parental involvement. In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems . 1-21. sourceNot checked yet

  Not checked in the free run.

• cites [13] [13] Griffin Dietz Smith, Siddhartha Prasad, Matt J Davidson, Leah Findlater, and RBenjamin Shapiro. 2024. ContextQ: Generated Questions to Support Meaningful Parent-Child Dialogue While Co-Reading. In Proceedings of the 23rd Annual ACM Interaction Design and Children Conference . 408-423. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The sociodemographic factors may have influenced the types of support families wanted from the robot.

“These may have influenced the types of support families wanted from the robot, which need to be addressed in future research.”

May combine two claims.

The study introduced PAiREd, an LLM-assisted interface integrated with an educational robot, enabling parents to review and revise LLM-generated learning content while adjusting their level of involvement by delegating roles between themselves and the robot.

“Additionally, we introduced PAiREd , an LLM-assisted interface integrated with an educational robot, enabling parents to review and revise LLMgenerated learning content while adjusting their level of involvement by delegating roles between themselves and the robot.”

May combine two claims.

The study was an in-home study with 20 parent-child pairs, each with a child aged 3-5 years old.

“In this paper, we conducted an in-home study with 20 parent-child pairs, each with a child aged 3-5 years old, to explore parental involvement scenarios, their use of LLM-generated learning content, and collaboration patterns with an educational robot in varying contexts.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Children from higher-income and more highly educated families tend to perform better academically.

“Children from higher-income and more highly educated families tend to perform better academically [67].”

• cites [67] [67] Selcuk R Sirin. 2005. Socioeconomic status and academic achievement: A metaanalytic review of research. Review of educational research 75, 3 (2005), 417-453. - sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The study developed a card-based activity kit, SET, to systematically map out parental involvement scenarios.

“We developed a card-based activity kit, SET , to systematically map out parental involvement scenarios.”