Over the past few weeks, journal editors, researchers and publishers have debated the appropriateness of listing the AI chatbot ChatGPT as an author. ChatGPT can generate convincing sentences by imitating the huge text it was trained on, and it can write abstracts of papers that can fool even professional researchers. But journal publishers all agree that it is inappropriate to list AI such as ChatGPT as a signed author, because they cannot be responsible for the content and integrity of the paper. Holden Thorp, editor-in-chief of the Science journal, updated the editorial policy on Thursday to prohibit the use of ChatGPT-generated text and not allow ChatGPT to be listed as an author. Journals require authors to sign a form declaring acceptance of responsibility for their work. Springer-Nature, which publishes journals such as Nature, also announced on Tuesday that ChatGPT cannot be listed as a signed author, but it allows authors to use tools such as ChatGPT in the paper preparation stage, as long as it is disclosed in the manuscript.
This article is transferred from: https://www.solidot.org/story?sid=73966
This site is only for collection, and the copyright belongs to the original author.