The growing homeschooling movement, now encompassing over 3.4 million American children and expanding rapidly across Canada and the United Kingdom, is finding new technological support through an AI-driven platform developed by researchers at the University of Notre Dame and partner institutions. Led by Mohammad Rashidujjaman Rifat, assistant professor of technology ethics and global affairs, the project introduces Homeroom, a large language model-based digital environment designed to empower parents as primary architects of personalized, culturally aligned education. The research, published for the Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems, directly addresses the limitations of conventional school-centered edtech, which typically prioritizes standardized pacing and classroom management over family-specific adaptability. Developed between April and July 2025, the platform was tested with Muslim homeschooling families in the greater Toronto area to evaluate usability, cultural responsiveness, and administrative efficiency. Homeroom architecture centers on three core objectives: positioning parents as active co-educators, enabling identity-affirming curriculum design, and fostering collaboration within closed, trusted peer networks. During trials, users leveraged the system to generate ethics-based storytelling, adjust learning materials to reflect specific heritage markers, and streamline record-keeping through automated portfolio generators and reusable rubrics. Participants consistently favored a trust-but-verify workflow, requiring full editorial control over tone, setting, and cultural motifs before materials were shared with students. Feedback also highlighted a critical gap in current AI alignment: generic value prompts often produced overly moralized or superficial content, indicating that precise, context-specific instructions are necessary for meaningful cultural representation. Beyond immediate utility, the study addresses foundational questions regarding human-AI collaboration in education. Rifat and co-authors Ayla Khan, Nouran Yahya, and Samar Sabie note that while AI personalization risks reinforcing ideological echo chambers, thoughtfully engineered systems can deliberately introduce cross-cultural exposure while respecting family boundaries. The platform design intentionally incorporates external perspectives to prevent isolation, balancing customized learning with broader community engagement. Researchers also observed that parents strongly preferred intimate, permission-based sharing circles over open forums, emphasizing the need for clear content attribution and flexible monetization options for original educational assets. The findings establish a replicable framework for developing edtech that scales personalization without compromising parental agency or cultural integrity. As homeschooling rates continue to outpace pre-pandemic trajectories, platforms like Homeroom signal a shift toward decentralized, community-driven learning ecosystems. Future iterations of the system will expand testing across diverse linguistic and socioeconomic groups, aiming to refine value-alignment mechanisms and reduce administrative friction for remote educators. The project underscores a broader industry pivot: moving from algorithmic standardization to adaptable, human-in-the-loop educational design that treats families as reflective partners rather than passive consumers.