The Luxury of Thinking

For years, we've argued that the future's true luxuries would be time, trust and truth. A fourth is arriving fast: thinking. Deep, independent, human-led thought is becoming scarce. Not because people can't think, but because it's become so easy not to.

USC researchers draw a parallel to the Industrial Revolution. Just as artisanal craftsmanship retreated to the margins as mass production took over, surviving as a luxury or a quiet act of resistance, independent thinking may follow the same path in an age of algorithmic abundance.

Cognitive sovereignty vs. cognitive atrophy

AI is amplifying a new cognitive divide, but it isn't simply between people who use AI and people who don't. It's between those who remain cognitively sovereign, owning their thinking, exercising independent judgment and resisting the pull of dependency, and those who begin to cognitively atrophy.

At the heart of that atrophy is the loss of productive struggle: the effortful, friction-filled process through which deep thinking and originality are built. Remove the friction, and you remove the workout.

What the research is showing

• We're starting to think alike. As more people rely on the same underlying AI systems, the diversity of human expression, reasoning and perspective is narrowing. Ideas filtered through the same tools converge, and collective intelligence suffers.
• We trust the machine over ourselves. People defer to AI over their own reasoning and intuition, even when the AI is wrong. Researchers at the University of Pennsylvania call this "cognitive surrender."
• Research is narrowing, too. A 2026 paper documents a scientific "monoculture": AI tools narrowing the questions researchers ask, the methods they use and the conclusions they reach. Similar convergence is showing up well beyond academia, including in the labor market.
• Confidence is eroding. In an American Psychological Association study, 58% of respondents said AI "did most of the thinking" in their work, and reported less confidence in their own reasoning and less ownership of their ideas.
• The gap is widening. Regular AI users scored lower on critical-thinking assessments, with the youngest participants (ages 17 to 25) scoring lowest. Those with deeper "cognitive reserves," built through years of education and mental effort, were better protected. AI may sharpen the already sharp while dulling those who lean on it as a crutch.
• The foundations are thinning. The OECD Skills Outlook 2025 finds that skill gaps tied to socioeconomic background are widest among the lowest achievers. And 40% of Americans didn't read a single book in 2025.

Memory, identity and the outsourced self

The stakes go beyond productivity. As Laurie Patton, president of the American Academy of Arts and Sciences, has observed, there are countless stories of people who lose their memory and, with it, their sense of self. Cognitive psychologist Barbara Oakley and colleagues make a related case in "The Memory Paradox": as external aids become more capable, our internal memory systems risk atrophy.

Previous technologies offloaded storage and retrieval. AI offloads the thinking itself. Cambridge researchers describe a shift toward "machine-mediated mnemonic processes," in which an AI may hold a fuller record of our lives than we do. When memory, expression and decision-making are increasingly handled for us, the felt sense of authoring our own thinking, and perhaps our own lives, begins to fade. Fluent, polished, generic output also crowds out the authentic voice through which people signal who they are and connect with one another.

Why it matters for all of us

AI has extraordinary potential to enhance collective intelligence by synthesizing knowledge at a scale no individual could. Deployed without intention, the same technology erodes the diversity of thought that makes collective intelligence meaningful.

That has consequences well beyond the workplace. A society that loses the habit of independent thought becomes more tribal, more susceptible to manipulation, and more responsive to "vibes" than to reasoned argument. Combined with the spread of mis-, dis-, mal- and digitally derived information, it weakens the shared foundations that good decisions, in organizations and in democracies, depend on.

What to do about it

The answer isn't to step away from AI. It's to use it without giving away the thinking that makes us human.

1. Protect the struggle (Agency). Decide where effort matters, and keep it there. Form your own view before you ask the tool for one, and treat AI as a sparring partner, not an oracle.
2. Question the first answer (Judgment). The most fluent answer isn't always the right one. Build the habit, individually and as a team, of asking what's missing, who disagrees, and what would change your mind.
3. Invest in deep work. Long-form reading, unhurried conversation and time to think aren't indulgences. Neuroscience shows they build the pathways for critical thinking, tolerance for ambiguity and complex reasoning. Make room for them, and reward them.

In a world of instant answers, the ability to think for yourself is becoming a competitive advantage. Like every luxury, it will belong to those who choose to invest in it.