British science magazine Nature has decided not to publish any images or videos created or enhanced using an AI generator like Midjourney. This decision was made after months of discussions and is ultimately about integrity and transparency. The lack of access to sources of data and images prevents verification, which is why no content created in this way will be published at least in the foreseeable future.
There is only one exception to this rule according to the editorial; if the content of articles explicitly talks about artificial intelligence, then content generated in this way is acceptable. Besides questions surrounding the transparency of training material used, the magazine also cites the fact that such AI tools only quote existing sources insufficiently or even incorrectly. Furthermore, the tools do not ensure that consent has been obtained to allow the use of content for training-making it easier for the magazine to reject them.
Although AI texts are partially allowed, there are certain conditions that need to be met. The generated texts must be documented comprehensively, and AI is not accepted as an author. The Science magazine allows AI-generated texts only with the express permission of the editors, and violation will be treated as plagiarism or altered images.
Nature believes that an AI revolution has great potential but could turn long-established academic conventions upside down. The magazine does not want to risk undoing all the achievements that have taken decades to create.
Until clearer legal requirements are in place, the magazine refuses to publish visual content created with AI generators. The move is a clear indication of Nature’s commitment to preserving the integrity of science and protecting authors from exploitation.
