Nature Journal Bans Use of AI-Generated Artwork in Posts

Last Wednesday, the prestigious scientific journal Nature announced that it will not post images or videos created with generative artificial intelligence tools. The ban comes amid the publication’s concerns about research integrity, consent, privacy and intellectual property protection, as the applications of this technology become more widespread in the world of science. and of art.

Founded in November 1869, Nature publishes peer-reviewed research from various academic disciplines, primarily science and technology. Is one of the most cited and influential scientific journals in the world.

AI-generated artwork may not be used in posts by Nature

Nature states that its recent decision on AI visuals came after months of intense discussion and consultation, motivated by the growing popularity and advancing capabilities of generative AI tools such as ChatGPT and Midjourney.

“Except for articles dealing specifically with artificial intelligence, Nature will not publish any content in which the photography, video or illustrations have been created in whole or in part using generative artificial intelligence, at least for the foreseeable future,” the magazine wrote in a text that it attributed to itself.

The publication considers the topic to fall within its ethical guidelines on integrity and transparency in its published works, and that includes being able to cite sources of information within images:

“Why do we ban the use of generative AI in visual content? Ultimately, it is a matter of integrity. The publishing process, for both science and art, is underpinned by a shared commitment to integrity. That includes transparency. As researchers, writers, and publishers, we all need to know the sources of data and images in order to verify their accuracy and veracity. Current generative AI tools do not give access to their source material for that verification to take place.”

Consequently, to all artists, filmmakers, illustrators and photographers who receive work commissioned by Nature “They will be asked to confirm that none of the works they submit have been generated or enriched using generative AI.”

Nature it also mentions that the practice of attributing existing work, a basic principle of science, constitutes another impediment to ethically using AI-generated works in a scientific journal. Attribution of AI-created artwork is difficult to do because artwork often comes from the synthesis of millions of images fed into an AI model.

This fact also leads to questions of consent and permission, especially in relation to personal identification or intellectual property rights. Also in this case, Nature He argues that generative AI falls short, routinely using copyrighted works for training without obtaining the necessary permissions. And then there’s the matter of counterfeiting: the publication cites the deepfakes as accelerators of the spread of false information.