Art is more than recombining elements: Why Nitin Nohria's essay on copyrighting gen AI outputs is flawed
- Sheelagh Caygill
- Jun 22
- 8 min read
If you are an artist or a creator, you will know that your work can be, or possibly once was, undervalued and under-appreciated
Now that we are in the age of artificial intelligence (AI), writers, graphic designers, illustrators, musicians, and other artists and creators are often undervalued, and the value of their work is frequently attacked. On and off the web, posts and articles argue for copyright laws to change; some proponents of AI want generative AI outputs to be recognized as art.
It's my role as a creative writer and co-founder of On Creative Writing to try to read and listen to as many of these articles, posts, and podcasts as I can to understand the arguments, key points, logic, and try and hazard a guess about what the near future may hold. Many of the articles and podcasts are well-written and make valuable points. Many more are unthinkingly enthusiastic about generative artificial intelligence, how highly valued it should be, and the spectacular things it will meet in near future. In doing so, they further contribute to the devaluing of creativity and the arts.
I can't respond to every piece. But when I read this essay by Nitin Nohria in MIT Technology Review titled AI copyright anxiety will hold back creativity, subtitle: Every original creation is an act of generative recombination. Why should the use of AI be held to a different standard? on art, copyrighting, and gen AI outputs I had to respond, partly because I disagree with the argument that gen AI outputs should be granted copyright, but also because the essay is flawed. Mr. Nohria is a successful, powerful, and influential man, and this essay is being shared and quoted across the web. He's the George F. Baker Jr. Professor at Harvard Business School and its former dean. He is also the chair of Thrive Capital, an early investor in several prominent AI firms, including OpenAI. He's written more than 40 books, and has been writing has been published in The Washington Post, The Atlantic, Harvard Business School, and other publications.
Art, copyrighting, and gen AI outputs
If you don't have time to read Mr. Nohria's essay 'AI copyright anxiety will hold back creativity', here's a summary:
The essay explores the relationship between human creativity and AI, arguing that AI’s generative recombination is not fundamentally different from the creative processes of renowned artists who drew inspiration from existing works. Mr. Nohria reflects on visits to museums where he observed how artists like Van Gogh, Dalí, and Picasso reinterpreted and built upon the work of others. He then draws a parallel to AI’s ability to generate content by recombining existing material. The essay contends that copyright law should not hold AI to a different standard than human artists and that meaningful human input should be the key criterion for copyright protection, regardless of the tools used.
Mr. Nohria argues that generative AI’s recombination of existing material is no different from the creative processes of renowned artists like Van Gogh, Dalí, and Picasso, and therefore, should not be held to a different standard regarding copyright and originality.
Let's start by pointing out the obvious. Mr. Nohria fails to mention that the "existing material" used in "AI's recombination" is protected by copyright, and use of this existing material is copyright infringement. Gen AI works on LLMs that use other people's work without their permission. That appears to be fine with Mr. Nohria. Would he feel fine if an author took sensitive intellectual copyrighted material from Thrive Capital, the company of which he is chair, to craft a financial thriller?
Nitin Nohria's essay 'AI copyright anxiety will hold back creativity' is built on flawed logic
The first flaw in Mr. Nohria's essay is that generative AI’s recombination of existing material is no different from the creative processes of renowned artists and so should not be held to a different standard regarding copyright and originality. This is simply incorrect. The creative processes of renowned and unknown artists are very different from AI’s recombination of existing material. In fact, it's incredibly bold to argue that generative AI’s recombination of existing material is no different from the creative processes of reknowned artists. Many artists say they do not understand their processes, or do not want to examine them, think about them, or discuss them. For many creators, there is a sacredness in what they do and in their way of seeing or ability to capture patterns. There is a mysteriousness in how they are able to crystallize an age or moment, and then rework, redraft, distill, until they are ready to offer their creation to the world.
Mr. Nohria's argument is further flawed because it doesn't understand artistic creativity, it mischaracterizes AI’s capabilities, and it neglects critical factors that distinguish human art from AI-generated content.
The essay’s comparison of AI to human artists is a false equivalence. Human creativity is driven by innumerable intentions and motivations, such as intention, purpose, and a desire to communicate ideas, emotions, or experiences. AI, in its current state, lacks this intentionality. Its outputs are generated based on algorithms and data, not on a conscious desire to create meaning or express a unique perspective. The essay also neglects the potential for harm and misuse associated with AI-generated content. AI can be used to create deep fakes, spread misinformation, and infringe on copyright, raising serious ethical concerns that the essay fails to address.
An incomplete understanding of artistic creativity and the combining existing elements
In 'AI copyright anxiety will hold back creativity' there is an incomplete understanding of what artistic creativity is. Mr. Nohria correctly notes that many artists combine existing elements. But he fails to recognize the way in which the combination of existing elements is done to make something unique, exceptional, or great. See my part two on this. Mr. Nohria also fails to recognize that this isn't always the case.
On the combination of existing elements, Mr. Nohria does not recognize the crucial role of originality, emotional expression, skill, and intentionality in human art. Artistic creativity involves not just recombination, to use his word, but also the infusion of new ideas, unique perspectives, and profound insights derived from personal experiences and a deep understanding of the world. Finally, he does not see, as Rollo May identified, the direct connection between the human artist and their subject matter, and how this is essential for genuine creativity, a connection that AI cannot replicate because it lacks consciousness and lived experience.
Misrepresenting the nature of AI generation
Mr. Nohria misrepresents what exactly AI generation is; it is a synthesis or even at times mutation - not genuine creation. AI algorithms operate by identifying and recombining patterns in existing data, without any understanding of the meaning or emotional content of that data. This process differs fundamentally from human creativity, which involves conscious thought, emotional investment, and a desire to communicate or express something meaningful. The essay’s analogy to Van Gogh, Dalí, and Picasso is misleading, as these artists transformed their source material through their unique artistic vision and technical skill, whereas AI merely regurgitates and remixes existing content.
No acknowledgement of hard work and time required to become a good creator
The essay overlooks the years of dedicated practice, skill development, and intellectual and emotional investment that human artists undertake to master their craft, and to me this is a surprising and shocking admission. Artists dedicate countless hours to honing their skills, developing their unique style, and gaining the knowledge and wisdom necessary to create meaningful and impactful work. This investment contributes significantly to the value and significance of their art. AI-generated content, on the other hand, requires no such effort or dedication, and therefore lacks the depth and nuance that comes from human skill and experience.
Oversimplified definition of art
Mr. Nohria's essay’s argument rests on an oversimplified definition of art. He assumes that if AI can generate something that resembles art, it is equivalent to human art. A deeper exploration into the philosophy of art will, in my view, reveal that human art is valued not just for its aesthetic qualities, but also for its:
Human element
Ability to move or persuade the reader, observer, listener, etc
Expression of personal identity, cultural heritage, or collective memory
Ability to convey complex thoughts, emotions, or philosophies
Potential to provoke, disrupt, or question the status quo
Power to create empathy and connection
Cultural context
Immediate message
Deeper and sometimes hidden meanings
Ambiguities
Power to heal, soothe, elevate and, at times, transcend.
Humans who appreciate and value art often want to know about the artist’s intention, the story behind the art, the work and research involved, the process, and more.
AI generated so-called art lacks these essential components.
Look out for false equivalencies in arguments from proponents of AI
False equivalencies are dangerous. They mislead people, distort facts, and stymie good discussions. They are rampant in the post-truth society in which we live. False equivalencies can be used to manipulate public opinion or deflect criticism. A false equivalence is often used deliberately to deceive people and undermine critical thinking and/or understanding by falsely equating things that are not actually equal.
In my experience, false equivalences are used often in arguments pushing for greater integration of AI in our society. Proponents of AI exaggerate or misrepresent the similarities between art and/or creative work and generative artificial intelligence, while ignoring or downplaying their crucial differences. It's like comparing human-carved, detailed stone masonry with a copy made on tracing paper, or a disposable camera with a high-end Hasselblad, or apples and oranges and claiming they are the same because they are both design/cameras/fruits.
False equivalence relies on faulty reasoning to create a misleading comparison
Ignoring key differences: False equivalence focuses on superficial or minor similarities while ignoring significant differences in nature, importance, or context.
Creating a false balance: This fallacy can create a false sense of balance or fairness, suggesting that both sides of an argument are equally valid when they are not.
The world is full of false equivalences. No wonder we're confused about so many things. We've seen them in advertising for decades, with an ad claiming a product is "as good as the leading brand" without evidence of comparable quality. Politicians equate the negative actions of two different politicians without considering the severity or context of those actions. In science, dubious publications offer a fringe scientific theory as equally valid to mainstream scientific consensus, and this is often the stuff of conspiracy theories.
Creatives and artists urgently need to navigate the ethical and legal challenges posed by AI in the arts
Mr. Nohria's essay raises important questions about the nature of creativity and the impact of AI on the arts, but many of its arguments are flawed due to a deep lack of appreciation of the creative life, a misunderstanding of artistic creativity, a mischaracterization of AI’s capabilities, and a neglect of critical factors that distinguish human art from AI-generated content.
Artists and creatives deserve a much deeper and nuanced understanding of these issues. But this will not come from business or proponents of artificial intelligence. Instead, we must create and share this message widely if we are to navigate the complex ethical and legal challenges posed by AI in the arts.
I encourage you to share this piece, or find your own words to speak out about this issue. It's not going to go away.
Comments