Is AI Making Us Stupid? Exploring Cognitive Offloading, Memory and Learning in the AI Age

Is AI making us stupid? In this Science Weekly episode, Ian Sample and Maddie ( Madeleine Finlay ) explore how AI and constant access to digital tools might change the way our brains work. They discuss a MIT study where students wrote essays with a chatbot, a search engine, or solo, and found that technology use reduced brain connectivity and memory for the essays. The conversation broadens to the idea of cognitive offloading, how memory can shift to external devices, and whether difficult problem solving is essential for deep learning. The hosts weigh the risks and benefits for students and adults, and argue for smarter ways to design tasks that incentivize thinking rather than reflexive tool use. The takeaway: use AI to assist, but preserve opportunities for hard thinking and practice.

Overview: AI, the brain and cognitive offloading

Is AI making us stupid? This Science Weekly episode investigates a central question: as we increasingly rely on AI and other digital tools, will our brains become less capable of thinking deeply, creatively, and quickly? Ian Sample guides a discussion with his co-host Maddie about how technology shapes cognitive processes, memory, and learning. The conversation foregrounds the idea of cognitive offloading, where the brain delegates tasks to devices or software, potentially changing what information we remember and how we problem-solve.

The MIT study and cognitive offloading

Maddie describes a MIT study in which 54 students wrote three essays under three conditions: with a large language model chatbot (ChatGPT), with a search engine, or solely with their own minds. While EEGs tracked brain activity and essays were linguistically analyzed, researchers found that technology use reduced brain connectivity, especially in the ChatGPT group, and some participants could not recall details from their own essays later. The researchers dubbed this a cognitive debt, signaling a drop in neural, linguistic, and behavioral performance over months of using AI tools. “Using technology, so using the chatbot or the search engine, this basically lowered the brain connectivity,” Maddie explains, noting scepticism from some scholars about the strength of the claim.

“Using technology, so using the chatbot or the search engine, this basically lowered the brain connectivity” - Maddie

Memory, Google effect, and cognitive offloading

The discussion broadens to the Google effect and digital amnesia—relying on external devices can dull our memory for information we could otherwise recall. Yet, there is nuance: memory for raw facts may fade, but the ability to retrieve information when needed can persist, suggesting a shift in what we remember rather than a blanket loss of memory capacity. The guests bring in Sam Gilbert, a cognitive neuroscience professor at UCL, to explain that humans have long offloaded cognition to external tools, and this isn’t inherently bad; it simply changes how we allocate mental resources.

The human brain is very opportunistic. It will use whatever opportunities it can find to simplify tasks - Ian Sample

Desirable difficulty and learning

The discussion moves to desirable difficulty, a concept popularized by psychologist Robert Bjork. Maddie explains that challenging practice, such as recall testing instead of passive rereading, strengthens learning. The idea is that some effort is necessary to solidify understanding and build robust cognitive skills, even if technology makes some tasks easier in the moment. The pair reflect on how education systems can preserve core cognitive development while still leveraging AI to support learning.

The best way to remember something in an exam is to test your memory - Maddie

Foundations, skills, and the role of offloading

Ian argues that a foundation of knowledge in the brain remains important for holistic thinking and cross-disciplinary reasoning. While calculators and GPS can enhance performance, they should not replace the mental foundations needed for critical thinking, problem solving, and memory for essential concepts. The guests note that external aids can help certain populations (for example, some evidence suggests digital technology use is linked to lower dementia risk in midlife), but cautions exist about over-reliance and skill erosion in specific domains like navigation for hikers or memory for exam-relevant facts.

We need to be smarter in the way that we design our assignments so that we can stretch students and give them an incentive to use their brains and not just use tools - Ian Sample

Practical implications for life with AI

The conversation turns to practical takeaways: when should we offload cognitive work to AI, and when should we push ourselves to solve hard problems? Maddie shares her personal experience with calendars helping manage a busy schedule, and the discussion emphasizes that AI should augment rather than replace cognitive effort. They acknowledge that AI has real benefits—from navigation aids to memory support for older adults—while warning against a culture that gradually saps the joy of problem-solving and the deep understanding that comes from effortful thinking.

AI risks us forgetting the joy and pleasure of the struggle - Maddie

Conclusion: a balanced path forward

The episode closes with a cautious, nuanced message: the brain can adapt to external supports, and cognitive offloading can expand our capacities, but it should not erode our core thinking abilities. The onus is on educators, researchers, and individuals to design tasks and environments that incentivize thinking, memory, and robust numeracy while harnessing AI to handle repetitive chores. The overarching takeaway is not to shun AI, but to use it intelligently—preserving the cognitive workout that keeps minds sharp.

"The human brain is very opportunistic. It will use whatever opportunities it can find to simplify tasks" - Ian Sample

"Using technology, so using the chatbot or the search engine, this basically lowered the brain connectivity" - Maddie

"The best way to remember something in an exam is to test your memory" - Maddie

"We need to be smarter in the way that we design our assignments so that we can stretch students and give them an incentive to use their brains and not just use tools" - Ian Sample