OpenAI to start watermarking ChatGPT’s text
OpenAI is preparing to add an invisible signature to ChatGPT and Codex text for users in the European Union, allowing authorised organisations to check if content was AI-generated. It is expected to roll out to eligible users on all plans within weeks. The feature subtly changes word-selection probabilities to create a pattern detectable by specialised tools.
read more at Firstpost


