As the AIEP project nears the end of our first piloting phase in San Francisco, we reflect not only on what we have built but also on what this work will make possible going forward.
What began as an effort to support families navigating Individualized Education Programs (IEPs) has become a practical example of how artificial intelligence can be developed with communities. The AIEP tool represents a model for intentionally and transparently building technology in ways that expand people's ability to co-design the systems and solutions that shape their lives.
Over the course of the project, parent leaders, educators, designers, and researchers provided the necessary perspectives to build a comprehensive tool endorsed by parents in the San Francisco area.
The outcomes of this collaboration extend beyond a single tool. They point toward a model for public-purpose AI grounded in lived experience, shared learning, and accountability.
The outcomes of this collaboration extend beyond a single tool. They point toward a model for public-purpose AI grounded in lived experience, shared learning, and accountability.
Below, we highlight the key elements that emerged from this year-long collaboration between the Burnes Center for Social Change at Northeastern University and Innovate Public Schools.
“I Was Afraid I’d Miss Something Important”: Why Parents Needed a Different Tool
At the center of the project is a free, open-source AI tool designed to help families understand and navigate IEPs. The tool directly addresses the challenges many families experience, regardless of their home language or prior knowledge of special education systems: long, difficult-to-understand documents, language barriers that delay understanding, and limited opportunities to ask questions without feeling intimidated or rushed.
One parent told us that her child’s IEP was so long that “I simply could not read it page by page, even translated page by page.” She worried that if she didn’t read every word carefully, she might miss something essential.
One parent told us that her child’s IEP was so long that “I simply could not read it page by page, even translated page by page.” She worried that if she didn’t read every word carefully, she might miss something essential.
After using AIEP, she shared that seeing a clear list of accommodations and services made her feel “much more confident and reassured.”
Another parent described how language barriers compounded that fear: “Because my English is limited, I often felt lost reading my child’s IEP. After trying AIEP, the information felt clear and straightforward.”By summarizing IEPs and translating the summaries into multiple languages (English, Spanish, Mandarin, and Vietnamese thus far), simplifying specialized educational terminology, and summarizing key sections, the tool is designed to support, not replace, family advocacy.
The tool's ability to quickly summarize and translate key information from IEPs enables families to engage more meaningfully in the IEP process and in meetings.
Unlike general-purpose AI tools, it includes safeguards to protect student privacy, such as redacting personally identifiable information, and has been vetted by families, educators, specialists, and translators to ensure the accuracy and quality of both summaries and translations.
Together, these features support clearer communication among IEP team members and empower parents and guardians to better understand and advocate for the services their children are legally entitled to receive. To date, more than 150 parents have used the tool, offering critical feedback that has shaped its design and functionality.
Learning as a Benefit of Participation
Alongside the tool, the project produced a Civic AI course and curriculum, designed for non-technical audiences and offered at an at-your-own-pace rhythm. The course addresses a key insight that emerged early in the project: meaningful participation in AI design requires shared understanding.
Rather than assuming familiarity with AI concepts, the curriculum introduces foundational ideas, explores opportunities and limitations, and supports parents and caregivers in developing the confidence to engage critically and thoughtfully with AI tools.
As we developed the project, we also shared these core AI concepts with parent leaders through training sessions and custom-designed materials. Learning is not an add-on to the project; it is a core condition for participation.
Turning Co-Design Into a Playbook Others Can Use
One of the most significant outcomes is the community-centered AI development playbook. Designed for a wide audience interested in developing projects similar to ours, the playbook documents the methods, decisions, and lessons learned throughout the project.
The playbook shows how families' ideas became real features, how we designed sessions that respected people's time, and how we decided what to build first.
The playbook shows how families' ideas became real features, how we designed sessions that respected people's time, and how we decided what to build first.
The playbook offers principles and concrete examples that others can adapt and use in different contexts. It is intended as a resource for teams seeking to implement community engagement approaches in AI-powered technology.
From Practice to Theory: Academic Contributions
The project also contributes to academic knowledge through a forthcoming peer-reviewed research publication submitted to the Design Research Society Conference.
This work situates AIEP within broader conversations in design research around co-design practices, infrastructuring, and sociotechnical systems.
While the playbook is designed to be accessible and practice-oriented, the research paper offers a more formal analysis of the project's process and implications.
This Is Just the Beginning
The work is far from finished. As next steps, we aim to expand the project in California, pilot the tool with families in additional school districts, and support a broader network of parent leaders.
We also plan to evaluate the quality and usefulness of the AI-generated summaries, incorporating qualitative feedback from parents, IEP advocates, and experts through a more systematic assessment to ensure the tool continues to meet families' needs as it scales.
Our goal is to ensure the tool continues to meet families’ needs as it scales, without losing the clarity and trust that parents value most.
Our goal is to ensure the tool continues to meet families’ needs as it scales, without losing the clarity and trust that parents value most.
The tool, course, playbook, and research are intended to inform future projects, support new communities, and contribute to a growing body of practice around community-centered AI. Our hope is that this work serves not only families navigating IEPs, but also others seeking to design AI systems that are accountable, participatory, and rooted in real-world needs.
In the AIEP project, we believe technology should help close unjust gaps in social systems, especially by supporting access to public services and resources, and that it needs to be designed accordingly.
Designing with communities requires time, care, and a willingness to let community knowledge shape technical decisions.
Designing with communities requires time, care, and a willingness to let community knowledge shape technical decisions.
This can be uncomfortable territory for many subject matter experts in tech and design, as it requires stepping back from established practices and sharing decision-making power, building AI not just for communities, but with them, in ways that expand rather than constrain their agency.
Read more about the AIEP journey: