# Bruce Schneier’s ‘Work vs. Gym’ Test Decides When to Use AI > Source: > Published: 2026-07-24 13:29:46+00:00 **July 24, 2026, (Inside AI)** — Bruce Schneier, a public policy lecturer at the Harvard Kennedy School and the University of Toronto’s Munk School, has proposed a straightforward framework for deciding when to use AI: treat tasks as either “work” or “gym.” The distinction, borrowed from AI researcher Daniel Meissler, cuts through the hype and offers a practical lens for evaluating AI’s role in cognitive and creative labor. At its core, the idea is simple. If the goal is just to get something done—moving heavy boxes, generating a standard report—then AI is a forklift. But if the process itself builds strength, like lifting weights or writing to sharpen thinking, outsourcing it to a machine defeats the purpose. Schneier applies this to his own classroom, where he sees students turning in AI-generated policy memos that read well but lack logical depth. “The writing assignments I give my students are gym tasks, not work tasks,” Schneier writes. He assigns memos not because the world needs more of them, but because the struggle of drafting, editing, and arguing hones critical thinking. Without that mental exercise, he warns, those skills atrophy—a trend employers are already detecting. Schneier’s framework arrives as AI writing tools become ubiquitous. A 2025 study by researchers at Stanford and MIT found that 62% of college students used AI for assignments, yet only 18% believed it improved their learning outcomes. The disconnect highlights a core tension: AI can produce passable work instantly, but it short-circuits the cognitive gains that come from doing it yourself. The work-vs-gym analogy extends beyond academia. For creatives, it clarifies a painful shift. Historically, writers and artists got paid for both utilitarian “work” (instruction manuals, corporate logos) and expressive “gym” tasks (novels, fine art). Now, AI can handle the former, slashing demand for human creators. “For the first time in human history, we can separate out when we need writing as work and when we want writing as gym,” Schneier notes. “And if AI can do most of the work-type writing, society doesn’t need as many human writers.” This isn’t just a labor problem; it’s an identity crisis. Many fiction authors, for instance, once funded their craft through technical writing gigs. As AI absorbs those jobs, the economic floor for creative careers drops. Visual artists face a similar squeeze: most images we encounter daily are functional, not artistic. A [2024 National Bureau of Economic Research working paper](https://www.nber.org/papers/w32345) documented a 14% decline in demand for freelance illustrators within six months of image-generation tools going mainstream. Schneier is careful to note that the framework assumes AI is actually competent and secure. “There’s no point giving an AI something that it can’t do reliably,” he says, citing risks from errors, bias, and cyberattacks. Trustworthiness remains a prerequisite. But once that bar is cleared, the work-vs-gym test helps individuals and organizations decide not just what AI can do, but what it should do. Yet the line between work and gym is blurry and shifting. A task that feels like drudgery today—say, drafting a legal brief—might be where a junior lawyer learns to construct arguments. Schneier acknowledges this: “The line between work and gym will change in the future as we humans adapt ourselves to a world with these new intelligences.” For now, he argues, the distinction is clear enough to guide daily choices, from climbing stairs instead of riding an elevator to writing a first draft without a chatbot. Schneier’s essay also touches on an incentive problem. No one pays us to go to the gym; the benefits are subtle and long-term. Similarly, students feel pressure to use AI because peers do, and the payoff of better reasoning skills is invisible in the short run. “For my students, incremental improvements in their reasoning and writing are equally subtle,” he writes. This mirrors findings from behavioral science: immediate rewards often override delayed, abstract gains. The work-vs-gym concept aligns with broader research on skill acquisition. A 2023 meta-analysis in *Psychological Bulletin* confirmed that “desirable difficulties”—the friction of learning—improve long-term retention and transfer. When AI removes that friction, it can leave users dependent and less capable. Schneier’s classroom experience bears this out: he can spot AI-generated prose by its “catchy, plausible, grammatically perfect” surface that masks a lack of coherent argument. Looking ahead, Schneier sees a future where society must make deliberate choices about the value of art and human effort. He draws a parallel to portrait painters, whose market collapsed with photography. “Maybe this time we can make different, more deliberate, choices about the value of art in our society,” he suggests. The work-vs-gym lens, while not a policy prescription, offers a starting point for those conversations—and a personal rule of thumb for anyone navigating an AI-saturated world.