AI agents stumble on legal reasoning
A new benchmark tests AI agents on real court cases, finding they often pick wrong conclusions even with the right evidence.
A new benchmark tests AI agents on real court cases, finding they often pick wrong conclusions even with the right evidence.