Question Everything. See who's reporting it.

Anthropic Details Claude's Watermarking Method for Code

Source: TechCrunch

The story in brief

Anthropic published a blog post Friday detailing how it will watermark text generated by its chatbot Claude, focusing on areas with arbitrary word choices in code, such as comments. The company addressed how the watermarking will work, whether it can be hidden through editing, and its impact on code, as part of broader efforts to ensure transparency and compliance with regulations like the EU AI Act.

NEWXFree on Android and iPhone. Open in NEWX