ChatGPT is no longer allowed to (co-)write scientific studies

It would actually be very practical if you could use the artificial intelligence ChatGPT to write a scientific study. Because it is easy for them to collect a lot of data quickly. But how reliable are the written investigations?

As the news portal “Nature” now reports, four scientific studies had already designated ChatGPT as a co-author. But the editors of scientific journals and studies agree that artificial intelligence does not meet the criteria for authorship. The reason: You cannot take responsibility for the content.

“We wouldn’t allow an AI to appear in the byline of a study published by us,” Holden Thorp, editor-in-chief of the Science journal group in Washington, told Nature. AI-generated text that does not properly cite the sources is also not allowed. That could be considered plagiarism.

However, some editors believe that the fact that an AI was involved could be mentioned elsewhere.

But as researchers’ experiments have shown, ChatGPT is not at all suitable for writing highly technical studies. AI often makes statements that aren’t necessarily true, and asking the same question multiple times will get you different answers, notes Alex Zhavoronkov, CEO of Insilico Medicine, a Hong Kong-based pharmaceutical company.