Are AI Chatbots Making Us Lose Control of Our Brains?

The Quiet Cognitive Crisis No One Is Talking About

At SXSW London this year, amid the music showcases, film screenings, and seemingly endless panels about artificial intelligence, one conversation stood out above the rest. Psychologist Gloria Mark, a professor at the University of California, Irvine, who has spent three decades studying how humans interact with digital technologies, sat down to discuss something that should concern every person who has ever asked ChatGPT to write an email, summarize a report, or explain a concept they could have figured out on their own. Her message was clear, measured, and more than a little unsettling: the tools we are building to make our lives easier may be quietly dismantling our ability to think.

This is not a fringe concern from a technophobe. Mark has watched the arc of digital technology's impact on the human mind play out in real time, from the early days of widespread internet use to the smartphone era to now. Each wave brought new conveniences and new cognitive costs. The AI wave, she argues, may be the steepest one yet.

Attention Spans Are Already in Freefall

Before we even get to AI, it is worth understanding the baseline we are starting from. Mark's research produced one of the most widely cited — and widely alarming — statistics in modern psychology. In 2003, the average human attention span while working on a computer was approximately two and a half minutes before the person switched tasks or got distracted. By 2020, that number had collapsed to just 47 seconds.

Forty-seven seconds. That is less time than it takes to boil water, read a short paragraph, or even finish a sentence out loud at a comfortable pace. And crucially, this relentless task-switching is not without consequence. Mark's work directly links constant context-switching to elevated stress levels, reduced productivity, and a growing inability to engage deeply with any single idea or problem. We have become, in a very measurable sense, worse at paying attention — and we did that before AI chatbots entered the picture.

How AI Makes Our Brains Lazy

Now consider what happens when you layer AI tools on top of an already-fragmented attention landscape. When you ask ChatGPT to draft your email, summarize a dense research paper, or evaluate the pros and cons of a business decision, you are not simply saving time. You are bypassing a cognitive process that psychologists call "depth of processing."

Depth of processing is what happens when you genuinely wrestle with information — when you read something difficult, pause to question it, connect it to things you already know, and ultimately integrate it into your understanding. This effortful engagement is what allows us to actually learn, retain information, and develop the kind of nuanced, critical thinking that separates good judgment from shallow pattern-matching.

When AI does that work for us, we skip the struggle. And just like a muscle that goes unused, our cognitive capacities can atrophy. The concern is not that AI gives us wrong answers. The concern is that even when it gives us right ones, we may be gradually losing the ability to arrive at them ourselves.

The Outsourcing Problem Is Bigger Than Productivity

This matters far beyond the workplace. Critical thinking, the ability to evaluate evidence, weigh competing arguments, and resist manipulation, is foundational to functioning as a citizen, a parent, a friend, and a decision-maker in every domain of life. If we consistently outsource the hard thinking to algorithms, we are not just becoming less efficient workers. We are becoming more intellectually dependent and, potentially, more vulnerable to misinformation, manipulation, and poor judgment in high-stakes situations.

Even Emotional Intelligence Is at Risk

Perhaps the most unexpected dimension of Mark's concerns involves our emotional lives. The rise of AI companions — chatbots designed to simulate friendship, offer support, and provide the feeling of being heard — introduces a subtle but profound risk. Real human relationships are difficult. They require patience, empathy, the willingness to be misunderstood and to repair that misunderstanding, the effort of showing up consistently for another person. That friction is not a bug; it is a feature. It is precisely what develops emotional intelligence, resilience, and the capacity for genuine intimacy.

AI companions require none of that effort. They are endlessly patient, perfectly agreeable, and always available. Mark warns that if current trends continue unchecked, loneliness, purposelessness, and emotional decline will deepen rather than ease — even as more people have access to AI that simulates connection. The simulation, however convincing, does not build the skills that real connection requires.

The Solution Is Not to Quit Technology

It would be tempting to read all of this as an argument for throwing your phone into a river and canceling your ChatGPT subscription. That is not what Mark is recommending. She is not calling for abstinence. She is calling for intentionality.

The prescription is deceptively simple: do the hard things on purpose. Read the actual book instead of the AI summary. Navigate somewhere unfamiliar without GPS. Meet your friends in person rather than messaging them. Write the first draft yourself, even if it takes longer and comes out rougher. Sit with a difficult problem before asking an algorithm to solve it for you.

Effort Is the Point

This is the counterintuitive insight at the heart of Mark's work: the difficulty is not the obstacle to learning and growth. The difficulty is the mechanism. The harder the cognitive task, the deeper the processing, the greater the reward — in terms of memory, understanding, creativity, and emotional development. Every time we choose the easier, AI-assisted path, we are making a trade. Convenience now for capability later.

A Critical Moment for How We Use These Tools

We are at an early and genuinely consequential moment in the relationship between human cognition and artificial intelligence. The technology will only become more capable, more convenient, and more deeply woven into daily life. The question is not whether AI will be present in our lives — it will be — but how consciously and carefully we choose to use it.

Gloria Mark's decades of research offer a valuable frame: technology is not inherently good or bad for the brain, but the habits we build around it matter enormously. Attention, critical thinking, and emotional intelligence are not fixed traits. They are skills, and like all skills, they are built through practice and lost through neglect. The choice, for now at least, is still ours to make.