Stanford Used LLMs to Scan 9,623 Local Legal Codes, Finding Dozens Still Mandating Segregation
chrmanning · x · 2026-09-12
- Stanford HAI and RegLab built an LLM-assisted review pipeline over 9,623 jurisdictions' local laws (3 billion words), covering roughly 252 million Americans.
- They found dozens of jurisdictions still formally mandating racial segregation, poll taxes, and denial of women's voting rights — all long held unconstitutional — with citizenship-based occupational license discrimination the most prevalent category.
- The brief argues LLM-assisted statutory surveys can make legal codes legible again and target reform, a task infeasible for human lawyer teams at this scale.
More from Safety
- Princeton to host AI Alignment & Safety launch event on Oct 19 — canondetortugas · 2026-09-12
- New Mexico lawyer fined $5,000 for ChatGPT brief with fabricated police testimony — nordicinst · 2026-09-12
- Agent liability: wrapping a script in AI doesn't change who pays for the harm — gerardsans · 2026-09-12
- AIRO launches: a weekly-updated dashboard of AI risk forecasts from frontier models — soumitrashukla9 · 2026-09-12
- Report: Yemeni weapons cell uses Claude Code to run parallel missile software programs — teortaxesTex · 2026-09-12
- Intel's alexvoica: fixation on AI kill switches distorts the safety debate — alexvoica · 2026-09-12