AI chatbots reducing human knowledge: Study finds

Overreliance on artificial intelligence (AI) chatbots could lead to a decline in human intelligence, according to a new study.

The research, conducted by the Department of Computer Science at the University of Copenhagen, found that repeatedly relying on AI for answers could reduce human knowledge, potentially resulting in poorer reasoning and cognitive abilities.

The study examined 27 different large language models (LLMs), or AI chatbots.

Researchers found that the information provided by the best-performing AI chatbot was 19% less diverse than the information available through the Google search engine.

In simple terms, AI chatbots tend to provide similar answers and similar ideas, meaning users may receive a narrower range of information and perspectives.

The researchers believe this could prevent people from developing a broader understanding of the world and may contribute to a decline in levels of intelligence over the coming years compared with previous generations.

The study tested 155 topics across 27 AI chatbots.

Researchers used 200 written prompts to obtain more than 70 million responses from the models.

GPT-5 produced the most diverse responses among the AI models examined, but even its responses were 18.5% less diverse than the information available through Google Search.

The researchers said that when people repeatedly receive the same information from AI chatbots, these systems are not only changing the way people acquire knowledge but may also be limiting the range of information to which they have access.