AI reliance may be creating a growing cognitive debt, researchers warn

The article argues that the negative impact of artificial intelligence on cognition is already evident, even if it is not yet reflected in formal performance metrics. While the long-standing Flynn effect has plateaued or reversed in nations such as Norway, the United Kingdom and the United States, researchers attribute part of this trend to the externalization of memory to digital tools. A 2011 study by Columbia University psychologist Betsy Sparrow demonstrated that people remember less when they anticipate future searches, a finding reinforced by a 2024 Frontiers in Public Health meta-analysis showing a self-reinforcing cycle of reduced encoding and increased searching.

Preliminary 2025 MIT Media Lab data reveal that writers who rely on large language models exhibit diminished activity in brain networks associated with memory, creativity and attention, with 83 % unable to accurately quote their own text minutes later—a phenomenon labeled “cognitive debt.” BCG’s survey of top managers indicates that half of respondents already notice skill erosion within their firms, and more than 60 % expect it to become a serious business risk within three to five years. The piece warns that unchecked reliance on AI could trigger a collective decline in critical thinking, undermining organizational adaptability and resilience.