María is six years old and in her first year at an elementary school in San Francisco. Last year, she was diagnosed with phonological dyslexia, meaning she struggles to connect sounds to words and to break down words into their individual sounds. When María’s mother, Clara, received this diagnosis, it felt like a dense fog had settled around her. Clara works long hours, cares for other young children at home, and feels overwhelmed communicating with María’s teachers and school staff in English, a language she is not comfortable speaking or reading.
When Clara was invited to María’s Individualized Education Program (IEP) meeting, an essential step in determining the specialized services and support her daughter would receive, she felt immense pressure. The IEP document itself was very long, complicated, and written entirely in English. Clara did not feel confident asking questions, and when she requested a translation of the document, she was told the process would take months due to a shortage of certified translators in the district. Delaying the signing of the IEP, she was warned, could also delay María’s access to essential resources.
It was only later, through conversations with parent leaders and community advocates, that Clara realized the accommodations and support described in María’s IEP were insufficient for her daughter to reach her full potential and receive the quality education she deserved.
Clara and María’s experience mirrors that of countless other families. This challenge drives our mission and inspires the creation of the AI-EP project.
AIEP: Building the Tool Together
Understanding these realities deeply shapes the AI-EP initiative. That’s why, in partnership with Innovate Public Schools, we are co-designing a tool with hundreds of parents, which will ultimately reach over 1,000 parents, caregivers, educators, and community leaders across California and Massachusetts through AI training and exposure to the tool.
Powered by artificial intelligence (AI), this free, open-source platform aims to eliminate key barriers families face when engaging with IEPs. Typically, IEPs are lengthy – often more than fifty pages – and packed with specialized educational jargon, making them challenging to understand even for native English speakers. As Clara’s and María’s experience demonstrates, for families who speak languages other than English, these obstacles can be an enormous barrier, drastically reducing their ability to advocate effectively for their children’s education.
The AIEP tool is designed to empower families to better understand and engage with their child’s IEP. Caregivers can upload the IEP document(s) provided by their child’s school to the AIEP platform. The platform’s AI tools will then translate the plan into multiple languages (Spanish, Chinese, and Vietnamese initially), break down specialized jargon into plain language, and distill lengthy and complex content into accessible summaries. Families can use the AIEP platform to clarify information and easily navigate their child's educational plan. Personalized recommendations and practical checklists generated by the tool will further equip families to advocate effectively during critical meetings with teachers, therapists and representatives from the school and district.
Why Community-Centered AI is Essential
AI has enormous potential to address some of society’s most pressing problems. Yet far too often, like the development of many other technologies, AI solutions are created without meaningful participation from the very communities they aim to serve. This disconnect can lead to tools that overlook real-world needs, reinforce existing inequalities, and have negative unintended consequences.
Community-centered AI development represents a significant shift toward a safer and more equitable approach. This methodology prioritizes ongoing, meaningful participation from community members, particularly those directly impacted by the issues being addressed, throughout every stage of the AI development process. For example, we are unboxing the prompts used to train the tool’s AI model to summarize sections of the IEP, engaging experts (both experts by training and experts by lived experience) to identify the types of terms the model should look for when scraping the IEP. We are doing focus groups to make sure the summaries the tool is providing are accurate, and most importantly, useful to parents when having their IEP meetings. By integrating community perspectives into the heart of technological design, we ensure AI systems genuinely reflect and serve the needs, values, and aspirations of the people who will use them.
Community Leadership at the Forefront
Central to the success of the AIEP project is the proactive leadership and insight provided by community members themselves. Our initial focus group sessions were conducted on March 19 and April 2, 2025, and involved two critical community roles: Parent Leader Navigators and Parent Leader Canvassers.
-
Parent Leader Navigators will receive extensive training by the research and core team, mastering the AIEP platform to support and educate other families in San Francisco. Their ongoing participation in focus groups will ensure the tool remains responsive to real-world needs and easy to use.
-
Parent Leader Canvassers will complete foundational training and contribute crucial feedback through outreach and participation in focus groups. Their insights will ensure that diverse community perspectives shape the development process, keeping the tool accessible and relevant.
Additionally, we are doing sessions with special education teachers and conducting individualized demo sessions of the tool with subject matter experts (including researchers and community activists) to gain a comprehensive understanding of the challenges faced by different stakeholders, without decentering our main audience: parents and caregivers.
Our Commitment to Responsible Research and AI Education
Beyond the development of the AIEP tool, our team is dedicated to rigorous research aimed at advancing inclusive AI development and community engagement. Our research explores critical questions, such as
-
how best to educate community members about AI
-
how to establish efficient and respectful co-design processes,
-
how AI can support, rather than confuse, communities who struggle to navigate existing systems, especially public systems
As part of this project, we are creating a series of bilingual videos (in English and Spanish) designed to explain the possibilities and opportunities offered by AI tools, as well as key points of caution. These resources will help parents and caregivers, especially those with little experience interacting with AI tools, better understand how to effectively utilize them. Additionally, we will develop targeted training materials for our specific tool, designed to support our Parent Leaders as they introduce and disseminate the tool within their communities.
All insights generated throughout the project will be documented in a practical playbook, intended as a guide for future community-centered AI initiatives. We believe this project can serve as a blueprint for how practical, community-led design processes can empower individual families and model an ethical, inclusive approach to AI development applicable to education and other sectors facing similar challenges. The project is supported by the Chan Zuckerberg Initiative and the Burnes Center for Social Change’s AI for Impact program. To read more about the project please visit our website.
Sofia Bosch Gómez is an Assistant Professor in the Department of Art + Design at Northeastern’s College of Arts, Media, and Design and an affiliate faculty member at the Burnes Center for Social Change.
