Students & civic learning
A Georgia student used coding AI to make public records easier to examine.
Max Kornstein’s archive project connects student initiative, software skills and the responsibility to preserve evidence.
Coding assistance gives students a way to turn questions about public life into software projects. Learning to verify the records, explain the code and correct mistakes is part of building something worth sharing.
The need: turn curiosity into something a learner can investigate
A student who wants to understand a community issue needs ways to gather information, organize it and ask questions of it. Software can help with that process, but writing the first version of a tool is only part of the learning.
The student also needs to decide what a record means, whether it is complete and how confidently a conclusion follows. A searchable collection can look authoritative even when a field was extracted incorrectly or the source describes an allegation rather than an established fact.
Coding assistance creates an opportunity to teach those responsibilities through a project. The goal is a learner who can explain both how the application works and why its information deserves—or does not deserve—confidence.
From an idea to a public-records project
An August 28, 2026 OpenAI Academy profile describes high-school student Maximilian “Max” Kornstein of Milton, Georgia, building Campus Evidence Lab with coding assistance. The project organizes public records about reported campus discrimination. The profile says he began without prior coding experience and continued checking records and improving the methodology. It is a supplier-published account; American Super Intelligence has not independently audited the archive. Read the student's story.
The educational value is active inquiry: a learner builds something to investigate a question rather than only receiving an answer.
Start with a question and a manageable collection
For example, a class would build a searchable index of public local-history documents. Students would define the fields, preserve a link to each original and check a sample of extracted dates and names manually. Ambiguous entries would remain marked for review rather than being silently filled with a model’s guess.
Another project would be a guide to publicly announced community events. Students would learn that a useful application needs current information, clear ownership and a correction process. A polished interface would not compensate for an outdated date or inaccessible registration link.
A science club would take a different route: analyze a public dataset, inspect missing values and write a small tool to reproduce a chart. Students would compare the generated code with an independent calculation and explain why the chart supports their conclusion.
Teach publication as well as programming
Before sharing a project, students would review permissions, accessibility and the consequences of errors. Records about real people require particular care with attribution and context. Source material should remain available so another person can check the interpretation.
Maintenance is another useful lesson. Who updates the data, reviews corrections and notices when a dependency stops working? A project is more educational when learners experience that continuing responsibility and document a handover that someone else can follow.
A good project should withstand questions
Evaluation should ask whether a student can explain the code, trace a result to its original source and correct an error. A civic application would also need a method for updating records and handling corrections.
American Super Intelligence would give more students access to the models, instruction and supervised practice needed for building, research and critical thinking. It does not imply endorsement of the archive, its contents or its methodology, or a partnership with Kornstein.
How Liberty will widen access to building and understanding
LibertyGPT will provide guided help with a student’s first application. Liberty OSS will let schools and community labs go further by running models, inspecting their limitations and adapting approved project tools.
Nano and Small variants will offer starting points for local experiments where available hardware supports them. More demanding coding work will use appropriately provisioned services, with Liberty Frontier Code as the proposed hosted specialist track. The learning pathway will explain those choices and connect code generation to testing, source verification and responsible publication.
Reusable lessons and supervised practice can help educators offer these projects without requiring every family to arrange its own technical support. Liberty will contribute an accessible foundation for inquiry, while the learner’s own reasoning, evidence and ability to correct the work will remain the measure of success.