Homeroom: A Value-Aligned and Community-Centered Homeschooling Platform

Paper Title

Homeroom: A Value-Aligned and Community-Centered Homeschooling Platform

Publication Info

• Topic area: Human–AI collaboration in homeschooling technologies
• Keywords: homeschooling, large language models, value-sensitive design, human–AI collaboration, culturally responsive personalization, community support, reflective workflows, curriculum generation, story generation, trusted circles

Background and Problem

• Problem / challenge: Existing educational technologies often prioritize standardization and scalability, which misaligns with the ethical, cultural, and pedagogical commitments of homeschooling families. These tools lack support for reflective planning, culturally responsive value alignment, and trusted community collaboration.
• Significance: Homeschooling is growing in diversity and scale, with families requiring tools that align with their unique values, cultural contexts, and pedagogical goals. Addressing these gaps can empower families to take control of their children’s education while fostering community support.
• Motivation and related work: Prior research in HCI emphasizes human-centered AI, explainability, and teacher–AI complementarity, but most studies focus on classroom settings rather than the home. Homeschooling parents, who juggle caregiving and teaching, need tools that are transparent, adaptable, and culturally sensitive. Current tools fail to meet these needs, leaving gaps in reflective workflows, value alignment, and community support.

Solution

• Proposed approach: Homeroom, a homeschooling platform that integrates large language models (LLMs) to support parent-led, culturally aligned, and community-centered educational practices.
• Novelty:
  1. A plan-then-generate workflow for reflective human–AI collaboration, enabling parents to preview, edit, and align generated content with their values.
  2. Culturally responsive personalization that treats family values as editable “soft constraints” rather than fixed settings.
  3. Small-circle community features for safe resource sharing, collaboration, and mental health support.
• Procedure and key techniques:
  • Story generation: Parents set objectives, tone, and values, approve outlines, and generate editable stories with tracked changes and reflective prompts.
  • Curriculum generation: Parents create lesson plans aligned with local standards, with AI-assisted analysis for compliance and engagement.
  • Community support: Invite-only groups for resource sharing, emotional support, and collaboration, with features like commitment cycles, Q&A spaces, and event planning.
  • Safeguards: Human-in-the-loop workflows, value glossaries, and provenance tracking to minimize risks of hallucination, bias, and misinformation.

Results

• Concrete findings:
  • Parents valued Homeroom’s plan-then-generate workflow, editable drafts, and concise rationales for generated content.
  • Culturally responsive personalization allowed families to fill representation gaps and share heritage in age-appropriate ways.
  • Community features like small, trusted circles and structured sharing reduced emotional and logistical burdens.
• Advantage over baselines:
  • Homeroom provided more control, transparency, and cultural alignment than traditional homeschooling resources or generic educational tools.
  • Parents could tailor content to their children’s needs and values, fostering trust and engagement.
• Experiments / evaluation:
  • Conducted with 15 Muslim homeschooling parents in the Greater Toronto Area.
  • Methods included task-based walkthroughs and semi-structured interviews to evaluate usability, cultural alignment, and community features.
  • Participants highlighted the importance of inspectability, flexible workflows, and culturally grounded personalization.
• Limitations and future work:
  • Study focused on a specific cultural group and geographic region, limiting generalizability.
  • Short-term evaluation; no longitudinal deployment or direct measurement of learning outcomes.
  • Future work includes multi-community studies, child-inclusive research, and scaled evaluations of governance and moderation.

Summary

Homeroom is a homeschooling platform that integrates LLMs to support reflective human–AI collaboration, culturally responsive personalization, and community-driven learning. It enables parents to generate and edit value-aligned educational content, share resources in trusted circles, and reduce administrative burdens. A study with Muslim homeschooling parents demonstrated the platform’s ability to enhance parental agency, cultural fit, and collaboration. While the platform shows promise, future work is needed to generalize findings, evaluate long-term impacts, and refine safeguards against bias and misinformation.