Should Machine-Made Art Be Subject to Copyright Protection?
Technology Policy Brief #77 | By: Steve Piazza | December 23, 2022
Header photo taken from: Ryan Abbott
Follow us on our social media platforms above
Browse more technology policy briefs from the top dashboard
Photo taken from: Stability AI
Last February, the U.S. Copyright Office upheld a rule that artwork created by machines or animals cannot be protected under copyright law. Referring to the Copyright Act of 1976, the U.S. Copyright Review Board reasserted that only works by humans can be protected.
Creativity is a constitutionally protected endeavor as stated in Article I Section 8 of the U.S. Constitution. What’s in question here, however, is not whether something is “creative” or not, but whether the Artificial Intelligence (AI) itself that produces artwork deliberately created without human intervention can be granted the protection.
In this case, researcher Stephen Thaler had applied for protections on behalf of the AI that generated an image, but Thaler was unable to convince the three member panel that his “Creativity Machine” algorithm passed as a person.
Neil Netanel, UCLA Law Professor and copyright scholar states that copyright law is used by “market institutions to enhance the democratic character of civil society.” It helps maintain a “careful balance between exclusivity and access” to creative expression necessary to a civil society.
In other words, disputes over who deserves compensation for work and who has access to it are unavoidable, and therefore laws designed to provide protections are required.
Copyright, as the country’s founders also understood it, plays a vital role in the economic stability of a democracy.
This is not an alarmist view, but a pragmatic one, considering that history is full of instances where totalitarian governments have seized control of creative expression in one way or another. Protecting creative expression is simply protecting freedom.
Regarding AI image generators, certain text is entered into algorithms that search the web for images which are then blended together to create a new image. In and of itself, artificially generated art is not a bad practice. It is quite amazing what is being created these days.
Some argue that it’s not art because in the end it’s generated by machine. Others contend it is because ultimately it requires some human involvement along the way. Either way, the purpose of the law goes beyond that debate because it raises legal questions about authorship. At its core it’s more about protecting the relationship between creator and end user, or in effect, the marketplace.
Yet the language of the law is constantly called into question as new technologies arrive. Thaler’s attempts for protection (he’s been applying for similar protections, even suing in countries around the globe, and with some success) should be viewed as a wake up call to copyright lawyers and Congress that present day statutes in the U.S. are not enough to prevent any upset to the equilibrium.
Photo taken from: Selvam & Selvam: an Intellectual Property Law Firm
(click or tap to enlargen)
Entire industries have changed as a result of emerging digital content and delivery systems, and there have been casualties, particularly on the artist’s end. Providing protection directly to AI without anticipating and adapting to newly forming consequences will only continue the trend.
Copyright does not protect the ideas themselves, but only the expression of them. What makes the future of copyright protections even more complicated is that it’s difficult to distinguish between an idea and its expression in a virtual world.
The test of a creative work is that it’s based on a human idea and resulting creation, but an AI work is resulting from the output of a machine ultimately on its own. Yet, as Thaler would argue, that independence is the point. Just as some citizenship rights are granted to a corporation, authorship to machines logically follows.
As these practices are allowed, one tends to look at society in a completely different light. It begs the question: How secure is someone supposed to feel if some of their constitutional protections are weighted equally against abstractions?
It’s comforting to know that the Office of Copyright has recently been addressing the issue of new technologies. In 2021, it held a conference (along with the U.S Patent and Trademark Office) entitled Copyright Law and Machine Learning for AI: Where Are We and Where Are We Going?.
This resulted in some discussion on possible approaches to sensible safeguards, such as shared authorship, tax incentives, and distinguishable data protection. Yet there seems widespread agreement there that changing the Copyright Law at this time is not yet warranted because of the belief that it still provides the best balance for existing creative markets.
Continued attention is warranted, though. If left unclear, the ambiguities inherent to the nature of digital content development and the law could lead to inequities, and an undesirable civil imbalance the law was attempting to avoid in the first place.
Click or tap on resource URL to visit links where available
This is the panel’s written response to Thaler’s request for copyright protection, which also includes work by the AI: