A top researcher says a new divide is emerging in AI use — and most people are on the losing side

A significant shift in how artificial intelligence is utilized is creating a divide among users, according to Vivienne Ming, chief scientist at the Possibility Institute and founder of Socos Labs. In a recent discussion, she highlighted that there are two distinct groups emerging in the realm of AI usage: a small percentage who leverage AI to enhance their thinking and a larger group who rely on it to do their thinking for them. Ming emphasizes that the prevailing trend is one of substitution rather than enhancement. Many individuals are outsourcing their cognitive tasks to AI, which she believes contributes to a growing cognitive divide. As AI tools become commonplace in various sectors, including coding and analysis, experts like Ming caution that excessive reliance on such technology may impair independent thought and reasoning abilities. Recent incidents, such as the downtime of Anthropic's Claude, have illustrated these concerns. Developers reported difficulties in completing tasks that had become routine, indicating that their capabilities were diminished without AI assistance. To investigate the effects of AI on cognitive skills, Ming conducted an experiment involving 39 UC Berkeley students and 33 participants from the San Francisco Bay Area. They were tasked with predicting real-world events using Polymarket data, either independently or with the help of AI. The findings revealed that about 90% to 95% of participants fell into two categories: those who depended on AI to produce answers and those who utilized it to confirm their own ideas. In contrast, a minority, comprising roughly 5% to 10%, adopted an approach Ming refers to as the 'cyborgs.' These individuals engaged AI as a collaborator, sparking discussions, challenging assumptions, and driving their problem-solving processes forward. Ming describes this interaction as 'productive friction.' Rather than passively accepting AI's outputs, these users actively questioned the technology, seeking to understand potential flaws in their reasoning. This collaborative dynamic fosters what Ming calls 'hybrid intelligence,' which transcends simple human-machine interaction. Her research indicates that effective collaboration between humans and AI is less about advanced algorithms and more about inherent human qualities like curiosity and the ability to reason through uncertainty. However, she warns that the current trajectory of AI usage often leads to a decline in these very traits. Ming likens the situation to the overuse of GPS technology, which, while convenient, can erode navigational skills over time. The implications of this trend extend beyond personal cognitive health, affecting workplace dynamics as well. Organizations that prioritize speed and efficiency may inadvertently encourage employees to accept AI-generated results without critical evaluation. This could result in a homogenized output, which Ming describes as 'AI slop.' She cautions that even if AI-generated answers are accurate, their ubiquity diminishes their value and uniqueness.