Company OpenAI said that it will stop marking all texts generated using this tool with an invisible watermark. The company is responding to EU law, the so-called EU AI Act, which requires that content created through AI be identifiable using tools that can reliably verify whether the text was created with the help of AI or not. Images, videos and music tracks already contain this identifier and now all texts generated using it will also have it ChatGPTJust for clarification, it is important to note that ChatGPT will not be the first to introduce this protection, as it already offers it Cloude.
In the coming days and weeks, it will be ChatGPT and Codex to implement this rule for all generated texts, using textGrain technology. It is essentially impossible to identify it in any way, because it works on the principle that it statistically calculates how the model chooses words or parts of words and thanks to this it can detect whether the text is written by a human or by AI. ProThe problem is that if you have some text generated and then rewrite it using your own words, it is no longer possible to detect whether it is AI text or your own text. ProGoogle has also been fighting AI in an extreme way recently, having released a Spam Update in recent days that detects websites that produce so-called Thin content using SAFE protection and subsequently removes them from search and Google Discover. It seems that AI will remain in our lives forever, but the extent to which we can use it will be greatly regulated.
So if you were using now ChatGPT for creating texts, writing homework or anything similar, then you are now out of luck and it doesn't matter whether it was a school project or something you subsequently published online. In both cases, everything can be easily detected and penalized.
