ChatGPT’s coding agent, Codex, helped 15-year-old Maximilian Kornstein gather and organize roughly 150,000 public campus civil-rights records in a single day, expanding an archive that had held about 4,000.
Max, a high-school student in Milton, Georgia, started Campus Evidence Lab after encountering antisemitism at school and hearing other students use racial slurs. When students are targeted for who they are, evidence is often scattered across police notices, court dockets, government investigations, and university statements, making public records difficult for reporters, researchers, or families to find. Max had never coded before he picked up Codex in February. He used it to build the first version of the archive in about two days, then to write the scrapers and filters that collect and organize material from sources including the Department of Education’s Office for Civil Rights, court dockets, and institutional responses. After an Emergent Ventures grant in July, the collection grew from roughly 4,000 records to about 150,000; Max says gathering the larger set took a single day.
The archive is group-agnostic, and includes harms directed at Black, Asian, Jewish and other students. Bringing those records together can show what was reported, how an institution responded, and where closer scrutiny may be warranted.
At Ohio University’s Athens campus, a police notification described an incident in which racial slurs and threats were allegedly directed at three Black students. At Haverford College, the Department of Education opened a Title VI investigation in August 2025 into alleged antisemitic harassment of Jewish students. An incident report or an opened investigation is not a finding; each gives the public a specific event, source, and institutional response it can examine.
Max says that during the Holocaust, his great-grandfather was taken from his home and killed. Antisemitism helped draw Max to this work; the archive’s scope reflects his conviction that students facing anti-Black racism, antisemitism, or any other form of discrimination deserve the same visibility.
Max is checking records, incorporating feedback on its methodology, and building tools that will make the material more reliable and easier for others to use.
