AI can already create art like pictures and videos, but there might be limitations to the creativity that it can currently generate.
Renowned author and thinker Malcolm Gladwell has waded into the debate on artificial intelligence and creativity, offering a nuanced perspective that distinguishes between mimicry and true innovation. For Gladwell, the essence of creativity lies in the unexpected and the discordant, a quality he believes current AI models are yet to replicate, even as they become proficient at generating content within established patterns.
Gladwell illustrates his point by referencing his work with the legendary musician Paul Simon. “I did a project with Paul Simon, the musician, and he has this lovely riff he would always give about how the ear is drawn to the discordant note,” Gladwell explains. “The thing that draws you into a melody is a little deviation. Weird, unexpected, bizarre sometimes, or a deviation from the conventional path that the listener thinks that they’re on. And so that’s what creativity is. It’s this discordant thing.”
This concept of creative deviation is, in Gladwell’s view, the current stumbling block for AI. “I don’t know whether I may be wrong, but that part of it doesn’t seem to be replicable with our current iterations of AI,” he says. To further elaborate, he deconstructs one of Paul Simon’s iconic songs, “Take Me to the Mardi Gras.”
“He took a falsetto singer, a reverend from Harlem, who was one of the great falsetto singers in gospel music, took him to Muscle Shoals, Alabama, which is the center of R&B, imported a marching band from New Orleans, and they got together and they did a Calypso song. So he combined four different musical traditions, and he was a white Jewish guy from New York. That’s one of his greatest songs.”
It is this complex fusion of disparate elements, this deliberate and imaginative juxtaposition of styles and traditions, that Gladwell argues is beyond the grasp of current AI. “AI’s not gonna give you that. ChatGPT’s not giving you ‘Take Me to the Mardi Gras,’ right?” he asserts. “It might give you a Taylor Swift song, like the fifth-best song on her album.”
Gladwell’s commentary cuts to the heart of a central question in the AI revolution: can a machine truly be an artist? While AI music generators like Suno and Udio are making headlines for their ability to produce entire tracks from text prompts, and even mimic the styles of famous artists, the debate over their creative capacity continues. These tools are trained on vast datasets of existing music, making them adept at pattern recognition and recombination. They can create a catchy, well-structured pop song because they have analyzed the components of countless others. However, Gladwell’s argument suggests that true artistry isn’t just about following the rules of a genre, but about knowing when and how to break them. The genius of Paul Simon’s “Take Me to the Mardi Gras” lies not in its adherence to a single formula, but in its audacious synthesis of multiple, seemingly unrelated, musical worlds. This act of creative rebellion, of finding harmony in the discordant, remains a profoundly human endeavor, suggesting that while AI may become an ever-more-powerful tool for creators, the role of the visionary artist is safe for now.