Anthropic introduces watermark in Claude's texts

Anthropic, an American company, has announced the introduction of a watermark in the texts produced by its language model, Claude. This measure is part of an effort to meet the requirements established by the European Commission under the AI Act, which aims to regulate artificial intelligence.

According to the company, the implementation of the watermark is part of its adherence to the Code of Practice under Article 50(2) of the European AI Regulation on Transparency. These actions aim to ensure compliance with new legislative standards related to artificial intelligence in Europe.