@lain@mirq They want to watermark the output directly by being deterministic every X amount of tokens. Probably just a way for feds to link LLM output to real people, because claude already wants ID in some scenarios I think.
@phnt@mirq > Anthropic has signed the EU AI Act's Article 50(2) Code of Practice on Transparency of AI-Generated Content, as a provider of both generative AI models and generative AI systems.
