Publishers had a problem with plagiarism, but the number of copied manuscripts was tiny compared to the hundreds of works they are receiving and which are created by platforms that use artificial intelligence (AI) such as ChatGPT or ChatSonic. During the first half of February, the number of spam cases received by the US magazine Clarkesworld doubled for the entire month of January. The publication has finished closing the reception of new works.

Clarkesworld, which publishes science fiction and fantasy stories, began to detect this practice between February and December 2020, with June of that year being the month in which more manuscripts had to be banned for being illegitimate works. During the years 2021 and 2022, some peaks were recorded. But the worst was yet to come: from October 2022 onwards, coinciding with the popularity of tools such as ChatGPT, the conversational chatbot developed by OpenAI.

To illustrate the barrage of falsified accounts that he has received, he has shared a graph in which you can see the increase in the growth curve of works that are plagiarized or are works written by bots.

According to data from Clarkesworld, the highest point of receipt of documents written by IA took place on Monday, February 20, when, of the total number of manuscripts collected in less than half a day, 10% were illegitimate and were not written by real authors. This month they have already exceeded the figure of five hundred works prohibited for having been generated by a machine.

Before, “these cases were easy to detect and rare enough that they were only a minor nuisance,” says magazine editor Neil Clarke. When one of these works was detected, he was expelled from future presentations. But now it is not so easy, because they recognize that it is difficult to locate all the new plagiarisms, although the person in charge has not wanted to detail how they know that the author is a bot.

Given the problem, Clarkesworld has announced that it has closed the reception of manuscripts, for the moment. In this sense, he is aware that “it is not a solution to the problem” and that, although he has some ideas to minimize it, “it will not disappear”, say those responsible.

Books written by AI are also abundant in Amazon’s Kindle store. As of mid-February, there were more than 200 eBooks in the Kindle Store that listed ChatGPT as author or co-author, including How to Write and Create Content Using ChatGPT, The Power of Homework, and the poetry collection Echoes of the Universe, according to reports. Reuters. And the number is increasing daily. There’s even a new subgenre on Amazon: books on using this chatbot, written entirely by ChatGPT.

The robot not only writes books under his name, but writes them for others. Hundreds of tutorials have sprung up on YouTube, TikTok, and Reddit demonstrating how to make a book in just a few hours and monetize that work in no time. A trend that worries the industry.

The AI ​​has also had problems with artists and cartoonists. Getty Images, one of the most important agencies in the world, has just taken legal action after understanding against Stable Diffusion, a machine learning model developed by Stability AI to generate high-quality digital images, on the grounds that it “infringed intellectual property rights , including copyright.

ChatGPT is a chatbot developed by OpenAI capable of almost anything. You can have a conversation, write an essay, solve a math problem, or develop software programming code. And all in a very natural way.

This tool is available for free and only requires an OpenAI account to use it, although there is a paid version, which offers subscribers more speed and depth in responses.

Its features are so advanced that OpenAI has just reached an agreement with Microsoft to integrate ChatGPT into Bing and improve search results, and incidentally compete with Google. In addition, some US media point out that the technology giant could integrate this tool into other products such as Word or PowerPoint.