Anthropic has announced that all AI models released after August 2 will include automatic watermarking for generated text and files, in compliance with the EU AI Act’s Transparency Code. This watermarking technology will also be extended to older models, ensuring that the watermark remains intact when text is copied and pasted, and may persist through some editing.
Watch for how the watermarking feature affects user behavior and content sharing. As AI-generated text becomes more prevalent, the effectiveness of this technology in ensuring transparency and accountability will be crucial, especially in educational and professional settings.