Two seemingly unrelated studies published in recent months point to a troubling pattern in how the brain responds to passive reliance on technology.
In one, Polish researchers found that colonoscopy doctors who regularly used AI assistance became measurably worse at detecting polyps when the AI was removed. In another, scientists in Singapore reported that babies exposed to screens before age two showed altered brain wave patterns that later predicted attention and problem-solving difficulties.
Together, the findings suggest that brains, whether developing or fully mature, can be reshaped when technology assumes the cognitive work.
The medical finding was unusually clear. After endoscopy centers in Poland introduced AI tools for polyp detection in late 2021, researchers examined what happened when doctors performed colonoscopies without assistance. The adenoma detection rate, a key measure of how well doctors identify precancerous growths, dropped from 28.4 percent to 22.4 percent after clinicians had been exposed to AI.
This was not a small sample or a marginal shift. Across 1,443 patients, doctors accustomed to AI support missed polyps they likely would have identified before. The implications extend beyond gastroenterology. One study on AI-related skill decay warned that the technology may accelerate deskilling without performers’ awareness.
The mechanism appears straightforward. When AI reliably flags suspicious tissue, human visual attention becomes less vigilant. Doctors scan less carefully, trusting the system to compensate. Over time, this creates a feedback loop in which the skill itself weakens.
A parallel pattern appears in early childhood development. Researchers at the National University of Singapore followed 437 children from infancy through elementary school. Babies who averaged more than two hours of daily screen time at 12 months showed measurable differences in brain wave patterns by 18 months, specifically elevated theta waves and higher theta-to-beta ratios in regions associated with attention and executive function.
These were not abstract markers. When the children reached age nine, they struggled with practical tasks such as remembering instructions, shifting attention, inhibiting impulses and managing multiple cognitive demands. Teachers reported more classroom attention problems and laboratory testing confirmed the difficulties were measurable.
The changes were not coincidental. Statistical modeling showed that altered brain wave patterns at 18 months accounted for nearly 40 percent of the link between infant screen exposure and later cognitive difficulties. The developing prefrontal cortex appeared to receive less of the stimulation required for normal development.
Both findings point to the same underlying issue, cognitive outsourcing. When screens provide stimulation for infants, the brain has fewer opportunities to develop top-down attention control. Passive visual input replaces active engagement such as exploration, object manipulation and interaction with caregivers. As the Singapore researchers noted, infants experience a
“video deficit”
meaning they struggle to learn from two-dimensional screens compared to real-world interaction.
A similar process occurs when AI performs detection tasks for doctors. The technology becomes a support that gradually weakens the perceptual skills it is meant to enhance.
What makes this difficult to detect is that neither group recognizes the change as it occurs. The Polish doctors were not intentionally inattentive. The decline in detection emerged without conscious awareness. Likewise, parents reported average infant screen exposure exceeding two hours per day, well above pediatric guidelines, likely without realizing the neurological consequences during a critical developmental period.
There is an irony in both cases. AI is introduced to improve medical outcomes and screens are often used to manage daily demands on parents. Yet the consequences appear later, when doctors must work without assistance or when children face tasks requiring sustained attention and cognitive flexibility.
The lead author of the colonoscopy study cautiously noted that continuous AI exposure
“might reduce”
performance without assistance. Despite the reserved language, the results were consistent. AI exposure was independently associated with lower unaided detection rates, even after adjusting for patient variables.
The infant screen time study also avoided definitive claims about causation, suggesting that screens may act as a
“proxy for the quality of parent-child interaction.”
Still, the relationship followed a clear dose pattern. Brain wave abnormalities increased as screen time rose from under one hour to more than four hours per day.
The comparison highlights a broader reality of neuroplasticity. Developing brains adapt to reduced cognitive demand by forming weaker executive networks. Skilled adult brains respond similarly when perceptual challenges are removed. In infants, the effects appeared persistent through age nine, with no evidence of full catch-up once academic demands increased. Each additional hour of daily screen exposure at 12 months was associated with measurable reductions in all three components of executive function nearly a decade later.
In adults, the decline occurred more quickly. Just three months of AI exposure produced statistically significant drops in unaided detection performance. This suggests that skill erosion can happen rapidly once cognitive effort is reduced, a finding with implications as AI expands into medicine, law, engineering and other fields.
Neither study argues for rejecting technology entirely. AI improves polyp detection when active and may prevent cancer. Screens are not uniformly harmful, particularly when used for interactive communication such as video calls with family.
The findings instead point to the need for deliberate limits. The Polish researchers suggested that doctors may require structured periods without AI assistance to maintain skill, similar to how pilots train for manual control despite automated systems. Pediatric guidelines already advise avoiding screen media before 18 months except for video chatting and limiting exposure afterward but adherence remains inconsistent.
Both studies illustrate how easily cognition can be offloaded without immediate warning signs. The doctor does not feel attention fading. The toddler does not register under-stimulation. The brain adapts quietly and the effects surface only when those capabilities are required.
AI systems and screens will become more integrated into daily life, despite the fact evidence suggests that maintaining cognitive effort is not optional. The brain depends on it.
Budzyń, K., Romańczyk, M., Kitala, D., Kołodziej, P., Bugajski, M., Adami, H. O., et al. (2025). Endoscopist deskilling risk after exposure to artificial intelligence in colonoscopy: a multicentre, observational study. The Lancet Gastroenterology & Hepatology, 10(10), 896–903. Correction published 11 September 2025.
Law, E. C., Han, M. X., Lai, Z., et al. (2023). Associations between infant screen use, electroencephalography markers, and cognitive outcomes. JAMA Pediatrics, 177(3), 311–318. https://doi.org/10.1001/jamapediatrics.2022.5674
