Skip to content

A journal will no longer receive manuscripts after registering hundreds of AI-created works

The American magazine Clarkesworld has been forced to stop receiving manuscripts after registering hundreds of works created by platforms that use artificial intelligence (AI) like ChatGPT or ChatSonic.

LOOK: Lima Major 2023: complete schedule and clashes of the group stage of the tournament

ChatGPT is a ‘chatbot’ developed by OpenAI that is trained to hold a conversation in testo. It is based on the GPT 3.5 language model and can offer natural responses, as if they were formulated by humans, as well as generate ideas, link them and remember previous conversations.

This platform is available for free and you only need to have an OpenAI account to use it. This detail worries cybersecurity researchers and artists due to the facilities that cybercriminals find to create malicious campaigns.

LOOK: Everything that is known about the iPhone 15: release date, price, models and more

This chatbot also does not receive the best reviews from creators and artists, who believe that other users can use this technology to create royalty-free works in which his style is plagiarized and commercialized.

Clarkesworld, which publishes science fiction and fantasy stories, has denounced receiving a large number of works written by an artificial intelligence of the type ChatGPT or ChatSonic.

LOOK: TikTok announces launch of API with the public data of its accounts for researchers

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 register plagiarism or are works written by ‘bots’.

Although it starts in 2019, the prohibitions begin to be noticed between February and December 2020, mainly due to plagiarism, June of that year being the month in which more manuscripts had to be prohibited because they were illegitimate works.

LOOK: Lima Major 2023: follow LIVE the debut of the Peruvian teams in the Dota 2 tournament

With several peaks during the years 2021 and 2022, the worst months have been from October 2022 onwards, coinciding with the popularity of tools such as ChatGPTwhen what they have begun to prohibit are works generated by AI.

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 500 works prohibited for having been generated by a machine.

LOOK: Netflix has removed the function that suggested titles randomly: why?

To alleviate this problem, Clarkesworld has announced that it has closed the receipt of manuscripts, a decision it has recently taken, ensuring that “it shouldn’t be hard to guess why”as he shared on Twitter.

The editorial has clarified that it has not closed the magazine it publishes, but has considered stopping receiving stories from authors, a solution that, at the moment, does not have a completion date. 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”.

“The detectors are not reliable. ‘Pay to send’ sacrifices too many legitimate authors. Printed shipments are not reliable for us”he commented, ensuring that the tools that allow these verifications “They are more expensive than magazines can afford.”

Source: Elcomercio

Share this article:
globalhappenings news.jpg
most popular