
Anthropic has revealed that it will soon watermark content that is processed ( not just generated! ) by any of its models. In a support article , Anthropic explained that it was rolling out machine-readable watermarks to comply with the…
View original source — Ars Technica ↗

