Why Relying on AI for Creativity Risks Long-Term Business Failure

AI’s Hidden Cost: Short-Term Gains vs. Long-Term Creative Erosion

Imagine assembling a team of all-star employees, investing heavily, and still finishing sixth in your division. That’s not a hypothetical scenario—it happened to Liverpool F.C. last season. According to MIT Sloan professor Sinan Aral, it’s also a powerful metaphor for how most organizations are deploying AI today.

Aral, a leading researcher on human-AI collaboration, has spent years running large-scale experiments on how humans and AI interact in real-world settings. His findings challenge conventional wisdom about AI integration in the workplace. “In about 85% of the studies we’ve seen,” he explains, “while adding AI to human beings improves human performance alone, most of the time it’s better to just let the AI do it alone.”

The Rational Fork in the Road

This data point—what Aral calls the ‘rational fork in the road’—poses a critical dilemma for leaders: If AI alone outperforms human-AI teams, the logical move seems to be replacing employees with automation. But Aral warns that this logic is flawed. The trap lies in prioritizing immediate productivity over sustainable creativity.

When ‘Good Enough’ Becomes a Creative Dead End

In one landmark study, Aral’s team randomized nearly 2,000 teams—some human-AI, others human-only—to create marketing ads for a real organization. The human-AI teams produced 50% more ads per worker with higher-quality text, seemingly a clear productivity win. Yet the ads began to look strikingly similar. “Ad copy starts sounding the same. Ad images start looking the same,” Aral notes. He labels this phenomenon “diversity collapse”—the gradual homogenization of creative output when AI, trained on the same publicly available internet, flattens the unique edges that define distinctive work.

The more teams delegated to AI, the more productive they became—and the more vulnerable they were to this collapse. What appeared as short-term gains masked a long-term erosion of creativity.

The Skills We’re Quietly Losing: Cognitive Offloading and De-Skilling

Aral’s latest research, published in his paper “The AI Augmentation Trap,” reveals an even more troubling trend: cognitive offloading—the act of outsourcing tasks you could do yourself—actively erodes the skills you’re handing off. Workers who rely heavily on AI for writing lose writing fluency. Junior employees de-skill faster than experienced ones, who retain professional reserves to preserve their capabilities.

“It leaves the worker worse off than if AI had never been adopted,” Aral warns. “In the long run, the productivity boost is real. So is the trap.”

Why Traditional Productivity Metrics Fail Creative Work

This dilemma reflects a deeper issue with how we measure productivity. The First Industrial Revolution’s legacy prioritizes speed, efficiency, and measurable output—an either/or model that overlooks the unseen work: the dormancy, marination, and synthesis that fuel truly original thinking. Aral’s research provides empirical evidence for what happens when we skip these critical phases.

What Leaders Should Do Instead: A Balanced Approach to AI

Aral emphasizes that avoiding AI isn’t a viable option—it’s one of the most disruptive technologies ever developed. The key, he argues, is to use AI as a catalyst for human creativity, not a replacement for it. Leaders must strike a balance: leverage AI for efficiency where it excels, but preserve the human judgment, intuition, and diversity of thought that drive innovation.

The alternative—full automation—risks not just creative stagnation but also the loss of the very skills that make organizations resilient in the long term.

“This is possibly the most disruptive technology ever developed.” — Sinan Aral, MIT Sloan School of Management