Anthropic announced the launch of a tool that detects novels written with artificial intelligence, explaining that the new “cloud” models will place an imperceptible watermark directly in the texts generated by artificial intelligence, which is the intended tool. The company explained that the watermark does not change the meaning of the text or its readability, but rather moves with it when it is copied and pasted. It is also a pattern or light text embedded in images or digital documents to help prove ownership and prevent copying or theft of the work, indicating that this step comes within Anthropic’s commitments to enhance transparency under the European Union’s artificial intelligence law. The Cloud models that were launched previously this month or later will support the tagging feature since their launch. Anthropic says: “It is working to add this feature to Older models.
