Anthropic to watermark Claude generated text, says quality will not be affected

CHENNAI: Anthropic has announced that future Claude models will use text watermarking to comply with the EU AI Act, while assuring users that the system will not affect the quality or content of Claude's responses.

Why is Anthropic introducing watermarking?

Anthropic said it is implementing watermarking to comply with the EU AI Act. Other major AI developers that have signed the same Code of Practice are also expected to implement watermarking.

Will watermarking affect Claude's output?

Anthropic said its watermarking method will have no practical impact on the quality or content of Claude's outputs. The difference between watermarked and unwatermarked text will not be distinguishable to readers.

Will anything be added to the text generated?

According to Anthropic, nothing will be added to the text and there will be no hidden characters. The watermarking will also not require additional tokens and will not make the service more expensive.

Can the watermark identify users?

Anthropic said the watermarks cannot be traced to a specific person, organisation or chat.