By all appearances, today’s AI language models are nothing short of miraculous. Whether it’s ChatGPT composing poetry, Perplexity generating essays, or Grok summarizing entire news articles, it’s easy to be awed by their linguistic prowess.

But beneath the polished surface of these AI systems lies a deep philosophical and ethical debate: are they truly intelligent — or are they merely repackaging human creativity in a sophisticated form of high-tech plagiarism?

In a recent episode of the Mind Matters podcast, guest host  Pat Flynn sat down with Dr. Robert J. Marks and Dr. Eric Holloway to explore this very question. Their conclusion? AI is impressive, but not intelligent — and certainly not creative.

Standing on the shoulders of giants — without climbing higher

Dr. Marks draws an important distinction. Isaac Newton (1642–1747 ), who gave us our laws of gravity, famously remarked that he was able to see further by “standing on the shoulders of giants.” Newton acknowledged the contributions of those who came before him formed that height. But it was Newton’s uniquely creative ability that made it higher.   

Modern AI, Marks argues, is a highly accessible mountain on which to stand. But it does not have the ability itself to make the mountain higher. Creativity belongs to humans.

Large language models like ChatGPT are trained on immense corpora of human text —books, articles, blogs, code repositories. They are brilliant mimics, not thinkers.  Marks notes they allow us to stand on the shoulders of giants, but any creativity is added by the human, not the machine.

The illusion of understanding

Image Credit: Narumol - Adobe Stock

Pat Flynn recounts his experience trying to use an AI model to research weight loss. The model confidently produced references that were wrong — sometimes blatantly so. It had to be repeatedly corrected before it got close to an accurate answer. This trial-and-error pattern exposes a key limitation: the model was generating plausible language, not demonstrating comprehension.

Why AI can’t be creative

The difference between imitation and creation becomes especially stark when we look at AI’s inability to improve itself through self-generated content. Holloway describes a problem called model collapse, where training AI models on content generated by other AI models leads to deterioration rather than progress. Unlike human artists, who refine their craft over time by engaging with new and challenging ideas, AI systems trained on themselves produce increasingly repetitive or meaningless output.

This mathematical limitation, according to Holloway, is strong evidence that AI will never exhibit true creativity. “[They] are completely parasitic on human creativity,” he says. AI can only remix what already exists.

The plagiarism problem

The legal ramifications of this “remixing” are already playing out in courtrooms. Major lawsuits have been filed against AI developers by The New York Times, Getty Images, and the Authors Guild — representing prominent writers like John Grisham and George R.R. Martin — for using copyrighted material to train AI models without permission.

Noam Chomsky has called this “high-tech plagiarism,” a label Marks and Holloway don’t dispute. When AI models reproduce styles, ideas, or even exact phrases from copyrighted works, the ethical boundaries become blurry. Is it inspiration, or theft? Courts will ultimately decide how intellectual property laws apply, but the underlying issue is clear: AI lacks originality because it cannot originate. It can only borrow.

Using AI responsibly

That doesn’t mean AI has no place in creative or academic work. Dr. Marks, for instance, uses ChatGPT to polish his writing. In engineering and science, AI can be a powerful tool to simplify complex tasks, solve equations, or explore data. But, as both guests emphasized, it should remain a tool — an assistant, not a replacement.

Flynn raises a parallel with chess. The best players today aren’t humans or computers alone, but human-computer teams. The same principle will likely apply across disciplines: those who are already skilled in their craft will benefit most from AI augmentation. But misuse — especially to shortcut the learning process — can erode real human development.

Human connection still matters

There’s also the question of meaning. AI-generated art may look compelling, but it often lacks emotional or cultural resonance. Holloway points out that audiences tend to value art because of the human stories and intentions behind it. A perfect AI replica of the Mona Lisa would never fetch the same value as da Vinci’s original — not because of the painting  itself, but because of the mind that created it.

Even the U.S. Copyright Office has recognized this distinction. A recent case allowed copyright protection for AI-generated art only when it involved significant human input — many iterative prompts, refinements, and direction. This sets an important precedent: creativity isn’t just about the final product, but the process and intention behind it.

The smoke Is out of the bottle

AI is here to stay, and the disruption it brings is real. But the fear that it will replace human creativity is misplaced. Rather than supplanting human minds, it amplifies them — or, when misused, undermines them. The real danger lies not in what AI can do, but in forgetting what only humans can do.

As Marks put it, the question isn’t whether AI can create, but whose creativity is it using? And if we value creativity, understanding, and meaning, we must remember: those remain uniquely, and wonderfully, human attributes.