For some time the internet has been passionate about ChatGPT, which poses some problems for which here is already a solution.
Whether on forums, streamers or even on YouTube, ChatGPT shows how impressive and amazing artificial intelligence can be. In particular in its “reasoning” and in the elaboration of texts. Because yes, for those who don’t know, it is an AI developed by OpenAI and which specializes in dialogue. The problem is that it can be used for bad purposes, such as returning a university assignment or spreading false information.
An ultimate weapon against ChatGPT
In this context, Edward Tian, a young man currently studying, has just developed a program which allows you to check very simply by entering a text, whether or not it was designed by an AI. In an analysis, you can see if a human or a machine is the author of a text. Thus, the more the sentences and the structure of the text appear random to GPTZero (the name of Edward Tian’s application) the more the application will be “surprised” and will offer a precise answer via a score.
To help you understand what it is, the man posted a very concrete example on his social network via an educational video:
I spent the New Year building GPTZero – an app that can quickly and efficiently detect if a write-up is ChatGPT or human.
here is a quick demo with john mcphee’s “frame of reference”
A miracle of technology, the longer the text embedded in the software, the more accurate the diagnosis will be as many symptoms as a doctor can detect. This type of software could have the right to rapid democratization in the future, especially in the university environment. Imagine, for example, a teacher scanning a text by his student to verify that he is indeed the worthy author. Moreover, the student specifies:
Since yesterday, several investment funds have slipped into my DMs on Twitter.
In short, a beautiful weapon that will allow in the future to live with AI in a healthier way.
