Anthropic clarified its AI text watermarking implementation using Google DeepMind's SynthID-Text standard, which embeds detectable word-pattern signatures without special characters. The watermark cannot be fully removed by partial edits, is invisible to humans, and doesn't impact output quality or processing costs. Future API tools will detect AI-generated content, addressing concerns about EU AI Act compliance affecting global users.