The US considers that the images generated by AI based on text do not have human authorship

The United States copyright office is clear about its position on images generated by artificial intelligence .


Artificial intelligence has  caused a real revolution in terms of creative processes for both texts and images. However, this has also caused a controversial issue to come up: the authorship of the photographs that are generated through these systems. Who should be considered the creator of such material? The AI ​​or the human that gave him the directions to get this result?

The US Copyright Office has been one of the first to weigh in on this contentious debate, noting that it believes text-based AI-generated images are not authored by humans The entity provided  a “style guide” where it clarifies that a photograph generated solely by a description does not qualify to have human authorship.

A debate on the authorship of the images generated by AI

In addition to the above, the office also clarified its position by comparing the AI ​​creative process to “instructions for a commissioned artist.” “An artist may modify the material originally generated by AI technology to such an extent that the modifications meet the standard for copyright protection,” the institution said, comparing the case specifically with a magazine editor hiring a photographer.

From what has been said, it is understood that, if there is a work sufficiently “visible” or appreciable to understand that the work in question is different from the original image generated by the AI , authorship could be granted to the human. However, it will be quite difficult to define when a work qualifies as “new” after its respective modifications.


View this post on Instagram


A post shared by Jos Avery Portraits (@averyseasonart)

A legal vacuum in which it is necessary to intervene

From the US copyright office, they point out that each of the cases in which platforms such as DALL-E and Midjourney are involved will have their own individual treatment and study.

“In the case of works containing AI- generated material, the Office will consider whether AI contributions are the result of ‘mechanical reproduction’ or rather than an author’s ‘own original mental conception, to which [the author ] gave visible form. The answer will depend on the circumstances, particularly how the AI ​​tool works and how it was used to create the final work .

Undoubtedly, this situation with AI and copyright is quite reminiscent of what happened with the selfie photograph that a monkey captured of itself using a human’s camera. David Slater was brought to trial by PETA because it considered that the animal should benefit from having been the one who captured the image. The US judges mentioned that copyright protection could not be applied to an ape.

Similarly, these rights may not apply to image generators even though we know there are companies behind them. In addition, it is important to remember that the deal for the “monkey selfie” was that Slater would donate 25% of the income generated by the image to organizations “dedicated to the protection of the welfare or habitat” of the animal, so something similar could happen with DALL-E or Midjourney .