NEW YORK, NY – Over the past year, AI tools like OpenAI's DALL-E have transformed how art is created, enabling anyone to generate complex images from simple text prompts. While this democratizes creativity, it risks reducing art to a mass-produced product, stripping away the emotion and meaning rooted in human effort. As AI takes over more creative roles, it threatens what we understand as true creativity. Research from Columbia Business School Professor Sheena Iyengar suggests that artists shouldn’t live in fear – showing that people across the board view art made by AI as less creative, less skilled, and less impressive than art created by humans.
In the study, Bias Against AI Art Can Enhance Perceptions of Human Creativity, Professor Sheena Iyengar and Columbia Business School doctoral students C. Blaine Horton Jr. '25 and Mike White '25 investigated how AI-labeled art influences public perceptions of human creativity, ingenuity, and talent. The study revealed that human-made art is rated as more skillful, creative, and valuable, especially when compared to AI-labeled works. Across their experiments, the researchers consistently found that participants viewed art labeled as human-made as having higher skill and higher perceived value than identical pieces labeled as AI-made. This challenges the idea that AI’s presence could harm human creativity, finding that inherent bias against AI art may enhance how human-made art is valued.
“AI tools are advancing rapidly, sparking fears that human creativity could lose its unique value,” said Sheena Iyengar, the S. T. Lee Professor of Business at Columbia Business School. "However, our findings reveal a fascinating nuance: when audiences compare AI-generated work to human creations, they often view human work as more impressive. This bias against AI stems from its perceived inability to convey the deep emotions we instinctively associate with art, reaffirming the special resonance of human creativity.”
The research team conducted a series of six controlled experiments with 2,965 participants from 2017 to 2023 to investigate biases against AI-made art, examining how attribution labels affected perceptions of skill, creativity, and market value across different scenarios. In one instance, participants viewed individual images labeled as either AI-made or human-made and rated them based on skillfulness and monetary value, revealing a bias against AI-labeled art, where it was perceived as less skillful and cheaper in comparison. In another version, participants rated two images, one labeled as AI-made and the other as human-made. This allowed the team to analyze how direct comparisons heightened appreciation for human creativity. In their final iteration, they assessed art where AI was used by a human to create it, labeled as human-AI collaborations. Collaboration labels slightly eased biases but didn’t match perceptions of purely human-made art, showing that while AI can mimic creativity, it can't fully replace the value of the human touch.
Additional Takeaways:
- Emotional Recognition for AI Art: Despite biases, people still appreciate AI art for its emotional and visual appeal, showing that aesthetic impact isn’t exclusive to human creation.
- Increased Perception of Human Contribution: Collaborative art (human + AI) is seen as having more human input when compared to purely AI-made work. This perception changes when compared to human-only creations, where the human contribution to the collaboration is viewed as lesser.
"Artists shouldn’t fear AI replacing their work or diminishing the value of human creativity," said Professor Iyengar. "AI can be a powerful tool, enabling human-AI collaborations to emerge as a distinct art form that elevates, rather than erases, the unique role of human talent."
To learn more about the cutting-edge research being conducted, please visit Columbia Business School.
###
