Schimbare importantă pentru ChatGPT în Europa: OpenAI introduce watermark în textele generate de inteligența artificială

OpenAI plans to introduce an invisible watermark in texts generated by ChatGPT and Codex for EU users, as part of compliance with European regulations. This technology, named 'textGrain', embeds a statistical signal that can be detected later to identify AI-generated content. The implementation will begin on October 5, with options for API clients worldwide, while EU users will see the watermark in the coming weeks. However, the detection system is not foolproof, as modifying text can significantly reduce detection rates. OpenAI emphasizes that the watermark does not determine authorship or legality of the content, nor does it guarantee accuracy. Initial access to the detection tool will be limited to approved researchers and organizations.