Recently, the widespread use of artificial intelligence tools in education has become an increasingly relevant topic. In particular, students writing essays, research papers, and other assignments using AI instead of writing them themselves is causing many debates. As a result, the issue of identifying AI-generated texts is becoming increasingly important.

ChatGPT usage rate among students

Globally: According to a survey conducted by the Digital Education Council in 2024, 86% of 3,839 undergraduate, master's, and doctoral students reported using artificial intelligence tools, including chatbots, in their studies. In the USA alone: According to a study conducted by the Pew Research Center in 2024, 26% of students aged 13-17 use ChatGPT to complete school assignments. According to a survey conducted by Intelligent.com in 2024, 37% of 588 college students are currently using ChatGPT.

In 2024, ChatGPT ranked fourth among the most visited websites in Uzbekistan. The platform received 42.73 million visits during the year. The majority of users are students.

How are students in Uzbekistan using ChatGPT?

According to the results of a survey conducted by "Avlod Media" in higher education institutions of Uzbekistan, 87% of students use ChatGPT. This is a very high figure, indicating the rapid popularization of artificial intelligence tools in the educational process.

Users utilize it for the following purposes

• Usage share (among ChatGPT users)
• Information search 60%
• Grammar checking 20%
• Summarizing 15%
• Paraphrasing 5%

Now it exposes what you've copied

ChatGPT, the artificial intelligence created by OpenAI, has recently introduced a new, yet unofficially announced function. According to experts at a company called Rumi, ChatGPT is adding hidden Unicode characters to its responses. These characters are not visible to the naked eye but can be detected in special text editing programs, such as Sublime Text, or online tools.

These markers are particularly prevalent in long, academic, or essay-like texts.

Only in English?

Currently, this hidden marking function is only activated for responses written in English. This may be due to technical limitations or an initial testing phase.

OpenAI has not yet made any official statement about this development. According to Rumi representatives, this is likely intentional: if users become too aware, they might try to remove such watermarks (as they can be easily deleted in ordinary text editors). Therefore, the implementation of this function is being carried out discreetly and cautiously.

A temporary measure or a new direction?

Rumi experts consider this a temporary measure by OpenAI. This is because combating such hidden markers is not particularly difficult. Perhaps this method will be abandoned in the future - just like a previously created low-accuracy service for detecting AI-written texts that was shut down without any warning.

Hidden markers could be a new step towards identifying artificial intelligence texts. However, it is not yet known how effective this approach will prove to be.

A temporary measure or a new direction?

Rumi experts consider this a temporary measure by OpenAI. This is because combating such hidden markers is not particularly difficult. Perhaps this method will be abandoned in the future - just like a previously created low-accuracy service for detecting AI-written texts that was shut down without any warning.

Conclusion

Today, 9 out of 10 students use ChatGPT. Only 15-17% of them said that they use it for specific educational tasks (searching for information, checking grammar, summarizing documents).

This suggests that a large proportion of students are using ChatGPT for various, sometimes unspecified or personal purposes. The new function may serve as an important impetus for young people to develop skills in how to use technologies and how to derive effective and honest benefits from them.