In the early summer of 2023, ChatGPT traffic dropped significantly and some analysts speculated that it was due to the fact that students stopped using AI reading tools for homework and exams after the end of the school year. In September, when classes resumed, ChatGPT traffic grew again, which only confirms this theory, writes Insider.
According to Similarweb analytics, traffic to ChatGPT’s website has grown significantly since the start of the school year.
On the one hand, this should suit OpenAI, as there will always be demand from high school and university students, and teenagers who get used to using AI will continue to use it in the future while working. On the other hand, such seasonality indicates limited scenarios for using ChatGPT in real life, which should be a concern for developers.
If most of ChatGPT’s growth comes from cheating students, it means that this technology, or at least this chatbot format, will not be the dominant computing platform in the future.
Interestingly, OpenAI itself proposes to use AI in education. At the end of August, the company released a teacher guide that demonstrates some ways to use ChatGPT in the classroom, including tips and lesson plans. This guide also contains answers to teachers’ questions about AI.
The most important one is whether OpenAI can help teachers determine whether homework was completed by ChatGPT or another AI model, or by students themselves. Unfortunately, the answer is no, it cannot. Moreover, existing tools for determining content authorship sometimes make outright mistakes, emphasizing that Shakespeare’s texts or the US Declaration of Independence were created by artificial intelligence.