Ai's productivity push: is efficiency killing creativity?
The relentless pursuit of optimization, fueled by artificial intelligence, is quietly eroding the very creative processes it promises to enhance. While tech giants tout ai as a productivity panacea, a growing body of evidence suggests we're trading genuine innovation for a veneer of efficiency – and potentially undermining the long-term value of our work.
The echo of peggy olson's struggle
Recall Peggy Olson’s painstaking process in the final season of Mad Men. A dozen restaurant visits, countless discarded drafts, a notebook filled with scribbles – a seemingly inefficient mess. Yet, it was through that very circuitous route, that embrace of failure, that she arrived at a truly groundbreaking advertising concept. This messy, iterative process, so fundamentally human, is increasingly at odds with the ai-driven push for streamlined workflows.
Microsoft proclaims ai agents will “optimize repetitive and mundane tasks,” while Mark Cuban predicts an explosion of creator productivity. Zoom CEO Eric Yuan and even Bill Gates have alluded to a drastically shortened workweek thanks to ai adoption. But the reality, as is often the case, is more nuanced.
Instead of liberation, many white-collar workers are finding themselves working longer hours, shouldering wider responsibilities, all under the guise of enhanced efficiency. A sobering MIT study from last summer revealed that a staggering 95% of ai pilot programs failed to demonstrate a measurable return on investment. The promise of “more meaningful” and challenging work appears to be, at least for now, a mirage.
The cognitive limits of human ingenuity
The core issue lies in the inherent limitations of human cognition. “There’s only a finite amount of creative thinking a person can do in a day,” explains Emily DeJeu, a professor at Carnegie Mellon’s Tepper School of Business. “That messy, trial-and-error process of sifting through bad ideas to find a good one doesn't lend itself to optimization.” The work of synthesizing multiple ideas, connecting disparate concepts, and producing original output is “cognitively very demanding – it takes time, and it requires space and bandwidth.” The notion that AI can readily assist with this is, according to DeJeu, “a bit fallacious.”
Amazon, Google, and Meta are pouring billions into the AI race. But are we so focused on the destination that we’re ignoring the value of the journey?
Consider the analogy of air traffic controllers, who operate on segmented schedules with breaks – a recognition of the cognitive toll of their demanding roles. Similarly, the rush to fill newly freed-up hours with increased productivity, thanks to AI assistance, may be counterproductive. Recent research from Harvard Business Review found that AI tools didn’t reduce workload; they intensified it, leading to longer hours and expanded job responsibilities.
Employees are coding “on autopilot,” creating a surge in code for engineers to review. Breaks are vanishing as workers integrate AI tools into every spare moment, blurring the lines between work and rest. The researchers warn of “cognitive fatigue, burnout, and impaired decision-making.”
Ben Armstrong, director of the Industrial Performance Center at MIT, points out a crucial oversight: “It can be a challenge to know which tasks, even if they’re tedious, are actually contributing to your learning or final workflow, and which are truly superfluous.” The seemingly mundane process of analyzing data, for example, can provide invaluable context and reveal hidden errors – insights that a purely algorithmic analysis might miss.
The paradox of productivity
Cal Newport, a Georgetown professor and author of Slow Productivity, argues that the current implementation of AI is often bypassing the very deep work it should be supporting. “Much of the AI use right now is less about accelerating or eliminating tasks that are in the way of doing valuable production work, and more about trying to reduce or avoid the maximum cognitive tension of thinking more intently.” Workers are generating imperfect drafts and editing them backward, avoiding the discomfort of starting fresh.
The blank page, once a daunting challenge, is now an enemy to be avoided. As Zoom’s blog puts it, “The secret to finishing your work is to do the work…When you get stuck on that blank page, the best thing you can do is fill it with something. Anything.” But this approach, as a recent MIT study showed, can lead to a dependence on AI, resulting in diminished performance at “neural, linguistic, and behavioral levels.”
Perhaps the most telling insight comes from Karim Adib, a PR manager at Search Atlas, who recently abandoned AI-generated idea prompts in favor of old-fashioned walks and library visits. “It’s about being bored, being with your own blank page, finding your own ideas,” he says. “Everyone else has access to ChatGPT, and everyone's slowly getting better at using it. So, if I and someone else are executing exactly the same way, that gives me zero advantage.”
The truth is, many of the issues driving long work hours aren't solvable by AI. They're problems of workplace culture. As Newport concludes, “Many of the things that keep us so busy and extend our long workdays aren’t problems that AI can solve; they’re problems that culture can solve.”
The pursuit of relentless efficiency may be inadvertently stifling the creative spark that drives genuine progress.