00:00Anthropic says it will watermark text generated by its AI models, including Claude, as new European transparency rules take effect.
00:08The technology will be embedded at the model level and is designed to remain with text even when users copy
00:14and paste it,
00:15although it remains unclear how easily the watermark can be removed through editing.
00:19The company is also applying watermarking to files using the C2PA standard.
00:24The move puts Anthropic among a growing group of AI companies responding to regulatory pressure and concerns over identifying machine
00:32-generated content.
Comments