Anthropic announced it will watermark text output from Claude to comply with Article 50 of the EU AI Act. The watermarks embed a pattern in word choices that is undetectable to readers but legible to those with a key. Critics, including John Gruber, argue the system compromises word precision. ZDNET notes the backlash highlights user over-reliance on AI tools and unrealistic expectations of corporate loyalty.
No score is assigned. Sources and their independence are shown in the citation chain below.