Claude (Anthropic) will label AI-generated content to combat disinformation

Anthropic has announced that Claude, its artificial intelligence system, will implement labels to differentiate AI-generated content from that created by humans.

These labels, called watermarks, will be applied to generated texts and images, aiming to help social platforms identify and combat misinformation. The images will use the C2PA standard, already adopted by companies like Adobe, OpenAI, and Google, while details on how the labels for texts will function remain unclear.

Although it is not specified when these labels will be implemented, the announcement suggests an application in the near future. This initiative comes in the context of the new European law on artificial intelligence, which came into effect on August 2, giving AI companies a four-month grace period to comply.