[Paper Review] A Systematic Review on the Potential of AI and ChatGPT for Parental Support and Child Well-Being

Mohsena Ashraf|arXiv (Cornell University)|May 23, 2024

Diverse Approaches in Healthcare and Education StudiesMedicine3 citations

TL;DR

This systematic review analyzes 27 studies on AI and ChatGPT for parental support and child well-being, categorizing them into three domains: supporting children with special needs, assisting pregnant and postpartum women, and addressing ethics and privacy. It finds that while ChatGPT shows promise in providing accessible, personalized guidance, most post-ChatGPT research is conceptual rather than empirical, with significant concerns over hallucinations, bias, and lack of rigorous validation, highlighting a need for ethical frameworks and human-in-the-loop integration.

ABSTRACT

This research article explores the potential of Artificial Intelligence (AI) and Chat Generative Pre-trained Transformers (ChatGPT) in revolutionizing the landscape of parental support and child care. With the advancement of generative AI, conversational agent technology, and Large Language Models, these tools can be a great support for parents to guide and assist their children. Recognizing the challenges faced by parents in navigating the complexities of child-rearing, this study seeks to explore the applications of AI, particularly leveraging the capabilities of ChatGPT, to provide valuable assistance and guidance. For this purpose, we perform a comprehensive examination of the existing 27 literature studies to investigate the potential of AI-driven platforms for offering advice tailored to the individual needs of the parents. We break down these research works into three domains based on their research criteria, and categorize each domain into two subgroups: pre-chatGPT era, and post-chatGPT era. At last, we investigate the significant dimensions of these literatures and present some valuable findings and future research directions.

Motivation & Objective

• To investigate the potential of AI and ChatGPT in enhancing parental support and child well-being across diverse parenting contexts.
• To categorize existing research into three domains: supporting children with special needs, aiding pregnant and postpartum women, and addressing ethics and privacy.
• To compare pre- and post-ChatGPT era research in terms of methodology, focus, and research communities involved.
• To identify research gaps, particularly in ethical and privacy considerations for AI use in parenting.
• To provide future research directions for responsible, effective integration of generative AI in family support systems.

Proposed method

• A systematic literature review was conducted on 27 peer-reviewed studies published between 2017 and 2023, selected based on relevance to AI, LLMs, and parenting.
• Studies were categorized into three domains: (1) special needs children, (2) pregnancy/postpartum support, and (3) ethics and privacy.
• Each domain was further subdivided into pre-ChatGPT and post-ChatGPT eras to analyze shifts in research focus and methodology.
• A feature-based analysis was performed on 33 common attributes across studies, including research type, participant recruitment, and ethical considerations.
• Content analysis and thematic coding were used to extract insights from qualitative studies, particularly in the post-ChatGPT era.
• The review assessed the quality of ChatGPT-generated responses in studies, noting issues like hallucinations, fabricated references, and lack of citations.

Experimental results

Research questions

• RQ1 How has the emergence of ChatGPT transformed the landscape of AI applications in parenting support and child well-being?
• RQ2 What are the key differences in research focus, methodology, and participant recruitment between pre- and post-ChatGPT eras in parenting-related AI research?
• RQ3 In what ways can AI, particularly large language models like ChatGPT, support parents of children with special needs, pregnant women, and new mothers?
• RQ4 What ethical and privacy challenges arise when parents and children interact with generative AI in sensitive developmental and health contexts?
• RQ5 What are the major research gaps in the current literature, particularly regarding empirical validation and ethical safeguards for AI in parenting?

Key findings

• Post-ChatGPT era research shows a significant shift from empirical to conceptual studies, with only 25% of post-ChatGPT studies involving human participants, indicating a methodological gap.
• In the domain of special needs children, 5 out of 11 studies were published post-ChatGPT, showing growing interest in LLMs for behavioral and educational support.
• For pregnancy and postpartum support, 6 out of 10 post-ChatGPT studies explored ChatGPT’s role in reducing postpartum depression through personalized guidance.
• Ethical concerns such as hallucinations, fabricated references, and lack of citations were consistently reported in studies using ChatGPT, with 100% of reviewed post-ChatGPT studies noting these issues.
• The research community has expanded beyond computer science, with increased contributions from medicine, OBGYN, and education fields post-ChatGPT, reflecting broader accessibility and interest.
• Despite its potential, ChatGPT cannot replace human interaction; instead, it should be integrated as a supplementary tool to enhance, not substitute, parent-child bonding.