kapynPolicy & Regulation

AI Models Are Watermarking Text—Will You Notice?

Anthropic will watermark all future Claude text outputs to comply with AI regulations. Google already watermarks Gemini output, and OpenAI plans to follow. The EU AI Act mandates watermarks for models released after August 2026, driving this rapid adoption. Critics argue text watermarking could degrade model quality since text offers less room for subtle modifications compared to images.

IEEE Spectrum·Sep 9, 2026

Opening Kapyn…