The company explained that the watermarking will create patterns in responses, undetectable to readers but detectable with a key, without impacting output quality. While light editing may not completely remove the watermark, a full rewrite would. For code, watermarks will be less prevalent, primarily affecting comments and having a negligible effect on the actual code produced. This initiative aims to comply with the EU AI Act's Transparency Code, which mandates the identification of AI-generated content.