Generative AI in Egyptian children’s everyday lives: Patterns of use, parental mediation, and cultural-religious negotiation in an Arab-Muslim context

Technical Summary: Generative AI in Egyptian Children's Everyday Lives

Problem Statement
While empirical research on children's engagement with Generative AI (GenAI) is expanding in North America, Europe, and East Asia, a significant gap exists regarding children in Arab-Muslim contexts. Existing literature is geographically narrow and often assumes culturally neutral media practices. However, children's media engagement is deeply shaped by language, religious values, family structures, and moral frameworks. Egypt, as the most populous Arab nation and a major producer of Arabic-language children's media, remains absent from the published record. The central problem is the lack of understanding regarding how Egyptian children (aged 8–12) interact with GenAI, how their parents mediate this use, and how Islamic religious frameworks and cultural values shape these interactions. Furthermore, established parental mediation theories, developed largely in Western, secular, nuclear-family contexts, may not adequately capture the strategies employed in extended-family, religiously observant Egyptian households.

Methodology
The study employs an explanatory sequential mixed-methods design conducted between February and June 2026 in Greater Cairo, Egypt.

• Quantitative Phase: A stratified purposive sampling strategy recruited 250 parent–child dyads (children aged 8–12). Stratification covered child age, school type (public, private national, private international), and district to ensure socioeconomic diversity.
  • Instrument: Parallel parent and child surveys were developed in Arabic, covering demographics, GenAI use patterns (frequency, tools, purposes), parental mediation strategies (adapted from validated scales plus new culturally specific items), child perceptions, and cultural/religious framing.
  • Analysis: Data were analyzed using IBM SPSS Statistics (v29) using descriptive statistics, Chi-square tests, ANOVA, and Exploratory Factor Analysis (EFA) to test the structure of mediation strategies.
• Qualitative Phase: This phase aimed to explain and deepen the survey findings.
  • Focus Groups: Eight parent–child focus groups were conducted with purposively selected families (stratified by use levels and mediation profiles). Separate groups for parents and children allowed for candid discussion.
  • Task-Based Observation: Thirty children from the focus groups completed short, age-appropriate prompts with a GenAI chatbot while observed by researchers and a parent. Children were asked to "think aloud."
  • Analysis: Audio recordings were transcribed in Egyptian Arabic and analyzed using reflexive thematic analysis (Braun & Clarke, 2019) with NVivo (v14). Triangulation was performed across survey, focus group, and observation data.

Key Contributions
The paper proposes three primary contributions:

1. Empirical: It provides the first large-scale, mixed-methods account of GenAI use among Egyptian children, addressing a critical gap in Global South literature.
2. Theoretical: It extends Parental Mediation Theory by proposing a culturally situated framework. The authors argue that the classical tripartite model (active, restrictive, co-use) requires extension to include:
   • Religiously framed restriction: Limits justified by Islamic ethical reasoning rather than generic safety.
   • Family-centred co-use: Mediation distributed across extended family (grandparents, siblings) rather than just the parent-child dyad.
   • Pre-emptive moral framing: Preparing children's interpretive resources before exposure to evaluate outputs, rather than reacting during or after.
3. Practical: It offers implications for educators, platform designers, and policymakers working with children in the Global South, emphasizing the need for culturally and religiously grounded interventions.

Anticipated Findings and Hypotheses
Note: As this is a study design paper (posted in 2026 with data collection described as completed but results presented as "anticipated" or "hypothesized" in the text structure), the summary reflects the specific predictions and expected outcomes outlined in the manuscript.

• Patterns of Use (RQ1): The study hypothesizes that a majority of children (8–12) have used GenAI tools (primarily chatbots and image generators) in the last three months, often without parental knowledge. Children are expected to report higher usage frequencies than parents. Use is predicted to vary by school type (highest in private international schools) and gender (boys favoring image generation/games; girls favoring conversation/creative writing).
• Mediation Strategies (RQ2): The study anticipates that factor analysis will reveal the three classical strategies plus the three culturally situated extensions. Restrictive mediation is expected to be the most common, but justified through moral/religious concerns rather than screen time. "Family-centred co-use" and "pre-emptive moral framing" are expected to emerge as distinct, empirically valid strategies.
• Cultural and Religious Framing (RQ3): Parents are expected to express concerns about the reliability of GenAI on religious rulings and moral advice. Children are hypothesized to use GenAI for religious queries (prayer, fasting, history). The study expects children to demonstrate a "layered ontological positioning," oscillating between viewing GenAI as a tool, a social interlocutor, and a morally bounded object depending on the context.

Significance and Claims
The paper claims significance in four areas:

1. Theoretical Extension: It challenges the assumption that parental mediation is a dyadic, nuclear-family phenomenon. By identifying "family-centred co-use," it suggests mediation theory must be reformulated for Global South contexts where extended-family socialization is the norm.
2. Temporal Shift in Mediation: The identification of "pre-emptive moral framing" represents a shift from reactive to anticipatory mediation, which is crucial for technologies (like GenAI) that produce unpredictable, non-replicable outputs that cannot be pre-screened.
3. Cultural Specificity: The study asserts that GenAI engagement is not culturally neutral. It argues that Egyptian children's interactions are organized by internal logics of Arabic-language identity and Islamic moral reasoning, which are not deviations from a universal norm but distinct instances of culturally situated media use.
4. Policy and Design: The findings suggest that current platform safeguards (often designed for Western, secular contexts) are insufficient. The paper calls for design priorities including meaningful age assurance, transparent limitations on religious queries, and parental controls that reflect extended-family structures.

Limitations
The authors acknowledge several limitations: the sample is limited to Greater Cairo and does not represent rural Upper Egypt; the cross-sectional design captures a single moment in a rapidly evolving field; self-report data may be subject to social desirability bias; and the "insider" positioning of the research team, while aiding rapport, carries a risk of unexamined cultural assumptions. Future work is recommended to include longitudinal designs and comparative studies across other Muslim-majority and non-Muslim Global South contexts.