Watermarking Diffusion Language Models

A new method for watermarking diffusion language models could improve AI content provenance and accountability.

Auto-scored: This is a real but partial technical step toward AI content traceability, with no direct risk escalation.