Anthropic Adds Invisible Watermarks to Claude

Anthropic updated Claude to include a hidden digital watermark in its AI responses. This tool helps people identify if a piece of text was written by a computer. The company says this is a necessary step for safety and transparency.

Some critics are concerned that these watermarks might subtly change how Claude writes. They worry the system could start choosing words based on a pattern rather than clarity or accuracy. If the AI becomes less helpful just to embed a code, the quality of its work could suffer.

Lawyers and tech experts are also weighing in on the legal side of things. They are debating how these marks will hold up in court or during copyright disputes. It is still unclear if these measures will actually stop bad actors or just create new technical hurdles.