Are students losing the habit of thinking?
One of the most interesting brain findings that emerged in the early 2000s was that veteran London cab drivers have significantly greater gray matter volume in their posterior hippocampus—an area of the brain that is associated with memory and spatial awareness. Neuroscientist Eleanor Maguire attributed the phenomenon to the difficult undertaking of memorizing a mental map of London streets, landmarks, and confusing routes that every cab driver needs to master in order to secure their license to operate.
This discovery became one of the most famous demonstrations of neuroplasticity, or the brain’s malleable ability to absorb new information and constantly evolve and adapt to manage day-to-day challenges. Neuroplasticity is activity-dependent, and scientists have often likened the brain to a muscle: the more we exercise certain neural pathways through effort and repetition, the stronger they become.
The reverse is also true. Any underutilized brain networks will eventually weaken and deteriorate over time. When particular skills or networks are no longer regularly used, the connections that support them can gradually weaken. This helps explain why we retain some abilities while losing proficiency in others. In the case of London cab drivers, for instance, research suggests that some of the changes in their brain may partially reverse once they retire.
Earlier this month, international news was abuzz with warnings from defense experts, computer scientists, and artificial intelligence (AI) specialists on how the creation of an artificial superintelligence can cause the demise of humankind. Evan Hubinger, Anthropic’s Alignment Science Lead, said that he personally believed there was a greater than 10 percent chance that AI could “kill all humans” within the next decade. As the world grapples with how to properly regulate increasingly powerful AI systems and prevent them from developing beyond our capacity to safely control them, there is another kind of harm whose consequences we are already beginning to feel in the present.
The Massachusetts Institute of Technology (MIT) recently released a report saying that AI is leading to what it called “cognitive surrender” among students. Cognitive surrender refers to how students, when faced with a challenging task, are choosing to immediately rely on AI chatbots rather than trying to accomplish it using their own intellectual agency. According to the report, students feel most tempted to do this when faced with a tight deadline.
The report warned that overreliance on AI has clear negative consequences on students. These include “diminishing critical thinking skills, weakening memory, eroding confidence, and undermining mastery.” Students seem to be increasingly skipping the cognitive process and patterns that are crucial for gaining deeper understanding and long-term retention.
This strengthens the concerns raised by previously conducted studies on how AI is potentially compromising an individual’s ability to learn. An MIT Media Lab study using EEGs to track brain activity during essay writing across different groups found that those who heavily relied on ChatGPT exhibited a drastic reduction in brain connectivity, particularly in areas responsible for memory load, semantic processing, and creative ideation. Meanwhile, those who wrote their essay without external tools had the strongest and most widely distributed brain connectivity.
Similarly, a large-scale study involving 26,811 Chinese secondary-school students found that after adopting generative AI, homework scores increased by 18 percent. However, their scores on monthly closed-book exams fell by 20 percent within six months. The largest learning losses were exhibited by students whose patterns of use suggested that they were outsourcing much of their homework to AI.
It is easy for educators to vilify the students for their lack of appreciation for the productive struggle involved in education. Perhaps an important starting point is how schools may have enabled an environment where students feel that the output matters more than learning itself. Some students are under immense pressure to stand out by accumulating credentials they believe are necessary to succeed. In the process, the deeper purpose of education can give way to a more performative one, where effort is optimized for grades, accolades, and getting into/graduating from the “right” school. If we have inadvertently conditioned students to believe that the grade on their assignment becomes more important than what the assignment was supposed to teach, then AI simply offers students an extremely effective way of giving us exactly what we asked for.
As neuroplasticity shows, part of the paradox of being human is that while we may naturally seek out what is safe, certain, and easy, our brains are actually wired to adapt and grow through hard work, challenge, and repetition. If the challenge AI poses to students is to resist the shortcut and engage in the more arduous process of thinking, creating, and problem-solving, then the challenge it poses to educators is to recalibrate our role in this new context so we can motivate them to choose the longer path. The role of schools has never simply been to deliver and transmit information. Educators are meant to help students make sense of the world, and to equip them with the skills, knowledge, and values to carve their place in it.
eleanor@shetalksasia.com
