Are we thinking ourselves out of a job? AI’s cognitive risks

Picture this: humans gliding through life in hover chairs, physically atrophied and mentally disengaged, dependent on screens for every interaction and robots for every task. Their days are spent consuming entertainment and products, never walking, never thinking deeply, never truly engaging with their world. When forced to stand, they stumble. When required to think independently, they falter.

This isn’t some dystopian fever dream. It’s the future Pixar imagined in WALL-E, and the thing is, they weren’t just making a cartoon. They were issuing a warning.

Fifteen years after WALL-E’s release, we’re already seeing early signs of the cognitive decline it depicted. A recent MIT Media Lab study tracking brain activity found that people using ChatGPT showed the lowest brain engagement and “consistently underperformed at neural, linguistic, and behavioral levels” compared to those working without AI assistance. The participants didn’t just perform worse—their brains were literally less active when AI did their thinking for them.

We’re facing a paradox that would have made excellent fodder for another Pixar film: the very technology promising to make us smarter may be making us cognitively weaker. And the stakes aren’t just individual careers or test scores. They’re about human cognitive resilience itself.

Use it or lose it

The concept isn’t new, but its implications have never been more urgent. Cognitive atrophy—the decay of mental abilities due to reduced use—operates on a simple principle that any gym-goer understands: if you don’t exercise your muscles, they weaken. The same holds true for your brain.

Memory, problem-solving, and critical thinking are cognitive muscles that need regular workouts to stay strong. When we delegate these functions to AI, we’re essentially putting our brains on a mental couch. Recent research has even coined a term for this phenomenon: AI Chatbot-Induced Cognitive Atrophy, or AICICA, which describes the specific ways AI tools contribute to cognitive decline.

The process unfolds through several mechanisms. First, there’s reduced cognitive effort. When AI provides instant answers, we lose the inclination to engage in deep, reflective thinking. Second, we develop what researchers call “cognitive laziness,” a decreased motivation to tackle mental challenges. Third, our working memory systems get less practice encoding and retrieving information. Finally, our analytical skills atrophy from lack of use, potentially reducing long-term cognitive resilience and flexibility.

The neurological evidence is striking. MIT researchers used EEG technology to monitor brain activity in people writing essays with ChatGPT versus those working without assistance. The ChatGPT users showed reduced alpha and beta connectivity, indicating serious under-engagement of the brain networks responsible for executive control and attention. Their brains were literally working less hard.

What’s worse, the decline happens faster than most people realize. Research across multiple domains shows that cognitive skills can begin deteriorating within weeks or months of reduced use. The “use it or lose it” principle applies ruthlessly to mental faculties. In our AI-accelerated world, the timeline for skill decay is compressing.

From cockpits to classrooms

To understand where we’re headed, aviation offers a sobering case study. Commercial pilots, among the most rigorously trained professionals in the world, are experiencing what researchers call the automation paradox. The systems designed to make flying safer are actually creating new dangers by eroding pilot expertise.

The numbers are stark: 77 percent of commercial pilots report that their manual flying skills have deteriorated due to heavy reliance on automated systems. Only 7 percent felt their skills had improved. Researcher Matthew Ebbatson found a direct correlation between a pilot’s manual flying competency and the amount of time they’d spent flying by hand. Skills decay “quite rapidly towards the fringes of ‘tolerable’ performance” without regular practice, with the most critical abilities—like airspeed control, essential for avoiding stalls—being particularly vulnerable.

The Federal Aviation Administration has become so concerned about declining pilot skills that it formally requested the International Civil Aviation Organization address the issue globally. “When automation ceases to work properly,” the FAA warned, “pilots who do not have sufficient manual control experience and proper training may be hesitant or not have enough skills to take control of the aircraft.”

The classroom tells a similar story. Students using GPS navigation systems provide a perfect example of how convenience breeds incapability. Research published in Scientific Reports found that people with greater lifetime GPS experience have significantly worse spatial memory when required to navigate without assistance. GPS promotes what researchers call “passive navigation,” you simply follow directions rather than building cognitive maps of your environment.

Eye-tracking studies reveal the mechanism: the more time people spend looking at GPS-like navigation aids, the less accurate spatial knowledge they acquire and the longer, more inefficient paths they travel when forced to navigate independently. Their brains, accustomed to outsourcing spatial reasoning, struggle to reconstruct basic wayfinding skills.

Perhaps most concerning is what’s happening in higher education. A systematic review of AI use among students found that 68.9 percent of those heavily relying on AI dialogue systems exhibited increased academic laziness, while 27.7 percent experienced measurable degradation in decision-making abilities. When MIT researchers had ChatGPT users try to rewrite their own essays without AI assistance, many couldn’t remember what they’d supposedly written, showing that the technology had bypassed rather than supported their memory and comprehension processes.

The pattern holds across professional settings. Digital technologies are creating what researchers term widespread “deskilling,” the atrophy of cognitive and social abilities that enhance human agency. From medical professionals who struggle without diagnostic AI to financial analysts dependent on algorithmic insights, the modern workforce is increasingly unable to function independently of their digital tools.

The enhancement trap

Here’s the cruel irony: AI genuinely can enhance human capabilities when used thoughtfully. Research shows that AI tools can improve learning outcomes by providing personalized instruction and immediate feedback, supporting skill acquisition and knowledge retention. The technology isn’t inherently harmful. It’s incredibly powerful with the potential to be used for both good and bad.

But that power creates a trust trap. As AI-generated solutions become increasingly accurate and dependable, humans develop what researchers call “blind trust in machine judgment.” We stop questioning, stop verifying, stop thinking critically about the information presented to us. Studies show that increased trust in AI-generated content leads directly to reduced independent verification of information, creating a feedback loop where our critical thinking skills atrophy from disuse.

The age factor makes this particularly concerning. Younger users, the so-called digital natives, show the strongest dependence on AI tools and score lowest on critical thinking assessments compared to older participants. Research consistently finds a negative correlation between frequent AI usage and critical thinking abilities, mediated by increased cognitive offloading.

The irony is stark: the generation most comfortable with AI may be the most vulnerable to its cognitive risks. Higher educational attainment provides some protection—people with advanced degrees maintain better reasoning abilities even when exposed to AI—but education alone isn’t sufficient armor against cognitive dependency.

In professional contexts, the challenge compounds. Industries from healthcare to law to finance increasingly rely on AI-generated insights, and while these tools can improve decision-making, they also risk creating professionals who struggle with independent analysis. When algorithms fail or produce biased recommendations, workers who’ve become dependent on AI assistance often lack the critical thinking skills to recognize and correct problems.

Building cognitive resilience in an AI world

The solution isn’t to abandon AI. That ship has sailed, and frankly, it shouldn’t return to port. The technology offers too many genuine benefits. Instead, we need to develop what might be called “cognitive resilience,” the ability to maintain and exercise our thinking skills while leveraging AI’s capabilities.

At the individual level, this starts with mindful AI use. Utilize AI for data processing, research, and repetitive tasks, but personally engage in critical thinking and decision-making. When ChatGPT gives you an answer, don’t just accept it. Question it, verify it, think through the reasoning. Use AI as a thinking partner, not a thinking replacement.

Regular cognitive exercise matters more than ever. Brain-challenging activities—puzzles, strategy games, learning new skills—aren’t just leisure pursuits; they’re cognitive maintenance. Continuous learning keeps your neural pathways active and adaptable. Physical exercise and meaningful social interactions support brain health in ways that no algorithm can replace.

For professionals, the key is developing what researchers call “AI-augmented humility,” recognizing that human expertise becomes most valuable when combined with AI rather than competing against it. This means understanding both AI’s capabilities and limitations, knowing when to trust the machine and when to override it.

Educational institutions face a particular challenge. Harvard Graduate School of Education researchers emphasize focusing on “intelligence augmentation” rather than replacement. Building human-AI partnerships rather than human-AI competition. This means teaching students to evaluate AI-generated content critically, understanding how AI works and where it fails.

Some schools are pioneering AI-resilient learning experiences that emphasize demonstrable skills and knowledge that can be meaningfully assessed. This might mean more oral exams, in-class written work, and collaborative problem-solving that can’t be easily automated. The goal isn’t to make AI-proof assignments but to ensure that learning happens in students’ brains, not just in their devices.

The aviation industry, having learned hard lessons about automation dependency, offers a model. Many airlines now mandate regular manual flying practice, even during routine flights. Pilots must demonstrate hands-on competency during training, not just systems management. Professional development programs focus on maintaining fundamental skills alongside technological proficiency.

Final thoughts

We stand at a cognitive crossroads. Down one path lies the WALL-E scenario where humans increasingly passive, dependent, and cognitively diminished. Our thinking skills atrophy while our devices grow smarter. We become, in effect, passengers in our own minds.

Down the other path lies a more complex but ultimately more human future. We learn to dance with AI rather than being replaced by it. We use technology to amplify our capabilities while maintaining the fundamental cognitive skills that make us human. We become, in the words of one researcher, “cognitively resilient.”

The research is clear about the risks. MIT’s brain studies show AI reducing neural engagement. Aviation data reveals skill decay from automation dependency. Educational research documents declining critical thinking among heavy AI users. GPS studies demonstrate spatial memory deterioration. The pattern repeats across domains: when we stop exercising our cognitive muscles, they weaken.

But the same research points toward solutions. Mindful AI use can preserve cognitive function. Regular mental exercise maintains thinking skills. Educational approaches that emphasize human-AI collaboration rather than replacement can produce learners who are both technologically fluent and intellectually independent.

The future isn’t predetermined. We’re not doomed to become the soft, dependent humans of WALL-E’s future. But avoiding that fate requires conscious effort, both individually and collectively. It means recognizing that cognitive health, like physical health, requires ongoing maintenance. It means understanding that convenience and capability aren’t the same thing.

AI will continue advancing. Will we advance alongside it? Maintaining our capacity for independent thought, critical analysis, and creative problem-solving? Will we use AI as a cognitive crutch or a cognitive catalyst?

The technology is powerful enough to enhance human intelligence or replace it. The choice, for now at least, remains ours. But like the cognitive skills themselves, that choice requires active engagement to preserve. Think of this as your brain’s daily workout—use it, or risk losing it.