Since we started talking about ChatGPT or even earlier, with the first reports of other AI models capable of generating texts made from simple queries, their more than likely use in many people’s minds is understandable, but not defensible. And it is so that in the hands of every student the temptation to resort to AI for the elaboration of works that should write the same it must be, it is easy to imagine, simply huge.
Those of us with gray hair, that is, those of us who had to resort to an encyclopedia at home or in the library to do the work we were asked to do in class, saw the advent of digital encyclopedias and later the Internet (with services such as Wikipedia, yes, but also with others that are much more specific, such as El Rincón del vago) how suddenly everything was much easier (in some cases too easy) and although we did not approve of some of these uses, specifically those based on sequence control+ca control+v keystrokes, we’d probably use them a lot too if they existed in our time.
With the advent of AI like ChatGPT, everything gets even more complicated, because unlike the options mentioned above, which always start from the same text, AI models generate different answers (albeit with the same sense and meaning) when faced with similar answers, which makes it somewhat difficult to detect texts with this origin. It’s true, yes, AIs have a very “personal” style (paradoxically, because they have nothing of a person) that can be identified, but of course prior training is required. In other words, today it is quite likely that many teachers are faced with texts that they do not know come from ChatGPT, but which students supply as their own.
However, this may change as OpenAI, the company behind the AI trend, created a tool that allows you to identify the texts created by ChatGPT and other similar text generation models. The tool, like ChatGPT, is completely free. To use it, you need to access it, which you can do at this link, and if you are not registered, follow the steps here.
Once you’re in, it’s very easy to use and even though the site is in English, you can use it with texts in Spanish. To run the test, copy the text you want to analyze into the box shown in it, taking into account that it must have minimum length 1,000 charactersand press the Submit button. A message will appear if the text appears to be generated by artificial intelligence
If OpenAI believes that the text is likely to be written by a person, it will display the following message:
And if it is not clear, it will indicate the following:
Keep in mind, yes, it is confirmed by its creators identification errors may occur, that if the texts have been modified, it is more difficult to determine their origin in AI and that the highest degree of reliability is offered by the texts in English. In order to test its operation and take screenshots of its responses, I tried it with several texts in Spanish (including the draft of this publication) and the results are… well, they leave a little aside. You have perfectly identified several texts written by ChatGPT, but there have been a few false positives and some doubt for texts of human origin.
Source: Muy Computer
Donald Salinas is an experienced automobile journalist and writer for Div Bracket. He brings his readers the latest news and developments from the world of automobiles, offering a unique and knowledgeable perspective on the latest trends and innovations in the automotive industry.