Anthropic announced plans to embed invisible watermarks in all text generated by its AI model Claude, beginning with new models released from August 2026. This update aims to improve content traceability and authenticity by incorporating digital signatures based on the C2PA standard.
According to The Decoder, these watermarks will be invisible to users but will serve as a reliable way to identify AI-generated content and verify its origin. The company will also sign files using the C2PA protocol, enhancing security and transparency in generated text.
This move comes as Japan’s markets increasingly scrutinize AI content authenticity and regulatory compliance, especially in sectors like FX and crypto where accurate information is critical for trading decisions.