Crosby and Nolla advance autonomous AI judges

The author argues that while AI may soon outperform human lawyers and doctors by operating within rule-based institutional systems, creating an autonomous AI judge presents much more complex alignment challenges. Unlike lawyers who have straightforward adversarial objectives, judges must interpret laws, weigh discretionary factors, and establish common law without a clear objective function. Consequently, implementing AI judges introduces significant risks regarding context, statutory interpretation, scheming, and sentencing. Open article →