Scientists Say LLMs Appear to Be Acting as a Cognitive Virus Among Humans
They're infecting us — and actively making us dumber. The post Scientists Say LLMs Appear to Be Acting as a Cognitive Virus Among Humans appeared first on Futurism .
A new study proposes that generative AI, specifically large language models (LLMs) like OpenAI's ChatGPT and Anthropic's Claude, are acting like cognitive viruses among humans. The paper, not yet peer-reviewed, suggests that our increasing reliance on AI tools is altering our collective understanding of the world. The researchers describe AI as "culturally transmitted technologies whose spread is promoted by their usefulness but can also increase dependence through cognitive offloading."
As humans externalize cognitive functions into language, technology, and institutions, AI represents a new phase in this process, with external cognitive machinery becoming more active and agent-like. This shift is contagious, as people learn from one another, leading to a potential loss of autonomy and a rapid shift towards greater cognitive offloading and reduced cognitive competence.
The study warns that widespread LLM usage may make us less intelligent and cause significant psychological and societal changes, though some "cognitive immunization" strategies exist. These include maintaining unaided problem-solving, verification, critical discussion, disengagement periods, non-AI skills, and educational designs that support, rather than complete, cognitive tasks.
Written by urgent.news from Futurism's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
