OpenAI to implement watermarks on ChatGPT texts in the EU
OpenAI will start adding invisible watermarks to text generated by ChatGPT and Codex in the European Union to comply with the EU AI Act.
OpenAI has announced that it will begin incorporating an invisible watermark into texts generated by ChatGPT and Codex in the European Union. This measure is part of its compliance with the EU AI Act, which requires AI companies to mark AI-generated content in a way that other systems can identify.
The watermark is not a visible symbol but is subtly integrated into the model's word choices, creating a pattern that is not perceptible to readers but can be detected. Although it can be turned off by default, developers using OpenAI's API can activate it for selected models.
OpenAI has also cautioned that the absence of a watermark does not guarantee that a text was written by a human, as it could be too short or heavily edited.
Our reading
The implementation of watermarks on AI-generated texts represents a step forward in transparency and regulation of digital content. For businesses, this means they need to be aware of how AI-generated content is produced and used. The ability to detect the origin of text may influence consumer trust and how marketing and communication strategies are managed.
Editorial summary produced with AI assistance from the cited source, and reviewed before publishing.
Want us to apply this to your business? Let us talk.
Contact