Generative AI has infiltrated almost every aspect of our daily lives, whether we like it or not.
It’s nearly impossible to avoid on social media, and it’s even showing up on dinner menus at the local diner. Some researchers have suggested that we’re already adapting our own behavior to suit the demands of AI tech.
AI’s wide proliferation shares some striking similarities with the ways viruses spread, as an international team of researchers argue in a new paper that has yet to be peer reviewed. They set out to lay out a new theoretical framework about how we’re increasingly becoming hooked on tools like OpenAI’s ChatGPT and Anthropic’s Claude, and how it’s changing our collective ways of understanding the world around us.
“Here we introduce the idea of [large language models] as cognitive viruses,” they write. “Culturally transmitted technologies whose spread is promoted by their usefulness but can also increase dependence through cognitive offloading.”
Researchers have previously warned that AI is actively damaging our cognitive abilities. Students, in particular, are choosing to use AI tools in the classroom, allowing them to offload much of their critical thinking.
“More generally, humans have always externalized parts of cognition into language, writing, institutions, and technology,” the researchers write. “This externalization has been the engine of rapid and accelerated change driven by culture. AI represents a new step in this process, one in which external cognitive machinery becomes increasingly active and agent-like.”
This transition to adapting LLMs also “has a contagion-like component,” because the spread occurs “because people learn from other people.”
It’s a slippery slope, the researchers argue, as the “loss of autonomy becomes self-reinforcing and the population can move rapidly toward a state of much stronger cognitive offloading and lower cognitive competence.”
Put simply, the widespread use of LLMs could be making us much dumber than we were before — not to mention wreaking complex psychological and sociological changes that we’re only starting to understand.
At the same time, there are ways to protect ourselves from being infected through a process the researchers dub “cognitive immunization.”
“At the population level, immunization does not mean preventing contact with AI,” the paper reads. “It means preserving the practices and institutions that keep human cognition active: unaided problem solving, verification, critical discussion, periods of deliberate disengagement, maintenance of non-AI skills, and educational designs in which the model supports rather than completes the cognitive task.”
More on cognitive offloading: Your Coworkers Are Becoming “Meat Proxies” for AI