Anthropic has introduced a watermark for AI-generated text that remains detectable even after edits, a move that appears to exceed the requirements of the European Union's Article 50(2). The EU law does not mandate marking AI outputs when the technology is used for standard editing or does not substantially alter the user's original content. Policy experts have criticized Anthropic's approach as overly broad, potentially affecting simple proofreading and translation tasks.
Related Articles
Don't miss out on breaking stories and in-depth articles.