OpenAI will watermark ChatGPT outputs by default—but only in the EU
Like other solutions, it is not especially reliable, and it's easy to circumvent.
This is a summary curated by AIFuture. Read the complete article at the original source:
Read the full story on Ars Technica