[ad_1]
The implementation of AI-based tools in our lives is a reality. Some of them like ChatGPT are reaching the training centers themselves, being used to write the assignments that teachers send to pass the subject. And this is causing even some teachers are returning to a traditional exam model with the objective that the students are the ones who work.
In addition, there is a lot of content that we have around us generated by AI without us knowing it. But the rise of this type of tools has also promoted the arrival of tools that allow you to detect texts that have been written by ChatGPT or if, on the contrary, a human has written it. But OpenAI believes that all these tools are not working properly and they are not trustworthy.
ZAO, the Chinese MOBILE APP that through DEEPFAKE turns you into DICAPRIO in SECONDS
Tools to detect ‘AI content’ do not work
OpenAI in the questions and answers section that are intended for teachers On their website, they practically state at the beginning that these detectors do not work. They are blunt when explaining that ‘in short, no’. They base this refusal on the fact that ‘no detector has been shown to reliably distinguish between AI-generated content and human-generated content.’
And this is something quite curious, since the very OpenAI has a tool which presents itself as an AI detector. Although now its own creators claim that It doesn’t work as expected. And that is why today there are no tools to know if a text has been generated with AI accurately. Obviously, this is a serious problem for all teachers.
In recent months, we have been able to see how true it is that these tools they don’t work properly. For example, we saw how an AI stated that The United States Constitution had not been generated by a human, but by an AI. And here we see that it can be really complicated to train a tool for this type of purpose.
And this well-known fact with the United States Constitution is recognized by OpenAI itself with the following phrase: “When we at OpenAI tried to train an AI-generated content detector, we found that it labeled human-written text like Shakespeare and the Declaration of Independence as AI-generated.” That is why it can be stated that There is still a long way to go to have adequate detection software.
Anyone can do this test on some of the AI content detectors on the web. In some cases you will be able to find that a text that you have written yourself claims that it has been written by the AI itself. Although also in the educational field if a few words are changed There will no longer be a way to identify what is generated by an AI.
Another option that may come to anyone’s mind is to use ChatGPT to detect AI-generated content. But the truth is that “ChatGPT has no knowledge of what content could be generated by AI. Sometimes it will have answers to questions like could this have been written by AI? But its answers are random,” according to OpenAI.
In Genbeta | Deleting chats in ChatGPT will not help you have more privacy. This option will
[ad_2]
Source link
