Visual works created by artificial intelligence without human input are not subject to copyright, says U.S. federal judge

This decision does not, however, concern all AI-generated images, but especially those created "without the tutelage of a human hand".

This decision does not, however, concern all AI-generated images, but especially those created "without the tutelage of a human hand".

A #visual #work #created by an #artificial #intelligence (AI) acting without #human #supervision cannot be protected by copyright in the United States, a US federal judge confirmed on Friday August 18. The judge was ruling on the validity of a previous verdict, handed down in early 2022 by another American body, the United States Copyright Office.

This Washington-based institution had refused to register in its repertoire of protected works an image generated by Dabus, an AI designed by Stephen Thaler. The scientist presents the image, entitled A Recent Entrance to Paradise, as an autonomous creation by Dabus: no human instructions had any influence on the final result. This radical approach could have come from a conceptual artist, but it was born in the mind of this researcher and founder of Imagination Engines, a company whose declared scientific objective is to build conscious, sensitive AIs. An AI should be recognized "as an author", argued Stephen Thaler's lawyer to the court.

Without the tutelage of a human hand.

In its ruling, the federal court recalls that American law has never "protected works generated by new technologies operating without the tutelage of a human hand", and that, according to well-established jurisprudence, it is human creativity that "grounds the possibility of receiving a copyright". In her view, the purpose of copyright law is to "encourage humans to engage" in creation.

However, this ruling does not condemn all AI-generated visual works. When it is used as a tool to assist creativity, the person who has employed it can, under certain conditions, succeed in obtaining copyright, if we rely on a notice from the US copyright office published in March 2023. At the time, the office argued that works in which the realization of an AI had been selected and arranged "in a sufficiently creative manner so that the work produced constitutes an original work of authorship" could be covered, on a case-by-case basis. While this opinion has no binding force, and comes fairly early in the emerging legal debate, it does seem consistent with a jurisprudence whose foundations were laid in the 19th century following the emergence of photography.

Neither this opinion nor the Columbia federal court ruling provides an answer to another question: can AIs be trained on the basis of millions of visual and textual creations culled from the Internet? This is another thorny debate that will undoubtedly take a few years to be definitively settled by the courts. Unless the political powers decide to set new rules for the game.