Washington, D.C.’s James Madison Memorial Building, which houses the U.S. Copyright Office. Photo Credit: UpstateNYer
Can you copyright works created entirely by AI? Definitely not, according to the U.S. Copyright Office, which has elaborated on this existing position and adjacent topics in a 50-page report.
That newly released report, specifically concerning AI copyrights, is the second that the USCO has put out in connection with an August 2023 artificial intelligence inquiry.
Unsurprisingly, the Office received more than a few responses (including from a variety of music industry companies, entities, stakeholders, and professionals) – to the tune of over “10,000 comments representing a broad range of perspectives.”
The inquiry’s first report, centering on soundalike and lookalike outputs, released last year. And an all-important third report, exploring “the training of AI models on copyrighted works, licensing considerations, and allocation of any liability,” is forthcoming.
Keeping the focus on the just-revealed copyrightability report, however, the USCO indicated that “[t]he vast majority of commenters agreed that existing law is adequate in this area and that material generated wholly by AI is not copyrightable.”
As many know, the position isn’t exactly new; the Copyright Office itself has already decided against protecting several works pumped out via AI. But what about works generated partially by AI and instances when AI tools (not straight generative models) are incorporated into the creative process?
The short answer is that it’s complicated – with the long answer proposing a continued case-by-case analysis of whether AI works can be copyrighted.
“Whether human contributions to AI-generated outputs are sufficient to constitute authorship must be analyzed on a case-by-case basis,” the Copyright Office summed up.
“Where AI merely assists an author in the creative process,” the Office proceeded, “its use does not change the copyrightability of the output. At the other extreme, if content is entirely generated by AI, it cannot be protected by copyright.”
Elsewhere in the all-encompassing AI copyrights report, the Office also tackled the possibility of protecting generative outputs owing to the individual’s role in penning the underlying prompts.
“The Office concludes that, given current generally available technology, prompts alone do not provide sufficient human control to make users of an AI system the authors of the output,” the USCO found.
“In theory,” the entity proceeded, “AI systems could someday allow users to exert so much control over how their expression is reflected in an output that the system’s contribution would become rote or mechanical. The evidence as to the operation of today’s AI systems indicates that this is not currently the case.”
Lastly, when it comes to apparently human-created works (like music or illustrations) inputted into and then modified by AI models – or largely human-created works containing well-defined AI components – copyright will solely protect the “perceptible human expressions” at hand.
“[The individual’s] own creative expression will be protected by copyright,” the USCO summarized, “with a scope analogous to that in a derivative work. Just as derivative work protection is limited to the material added by the later author, copyright in this type of AI-generated output would cover the perceptible human expression.”
And at least for the time being, it’d be wise to stick with the aforementioned case-by-case AI copyrights determinations as opposed to addressing the heretofore unseen issue via new laws, the Office reiterated in conclusion.
“Because the copyrightability inquiry requires analysis of each work and the context of its creation,” the USCO wrote, “statutory language would be limited in its ability to provide brighter lines. Unless and until future developments raise new problems, the Office does not recommend a change in the law.”