Research involving 1,923 adults finds stronger confidence in reasoning among users who actively challenge and refine AI output.
As generative AI becomes part of everyday work, new peer-reviewed research published by the American Psychological Association suggests the real story is not that AI weakens thinking, but that the benefits depend on how people use it. The study was spearheaded by globally renowned A.I. Neuroscientist, Sarah Baldeo, CEO of ID Quotient.
In a large mixed-methods study https://www.apa.org/pubs/journals/releases/tmb-tmb0000191.pdf of 1,923 workers across North America, participants used generative AI tools to complete 10 simulated work tasks involving planning, sequencing, ambiguous information, decision-making, and reflective reasoning.
The study found a clear distinction between passive and active use. Participants who modified, challenged, or rejected AI suggestions reported greater confidence in their own reasoning and a stronger sense of authorship than those who accepted AI output with minimal revision.
“AI can be an extraordinary cognitive tool,” said Baldeo, Executive MBA, PhD candidate in NeuroAI at Middlesex University, and author of the study. “The strongest outcomes were not linked to avoiding AI, but to staying cognitively engaged while using it. The users who questioned, refined, and overrode AI responses, reported greater confidence in their own thinking because of tool usage.”
Across tasks, 58% of participants agreed that “AI did most of the thinking,” highlighting how common passive use has become. But the findings do not support claims that AI causes cognitive decline, reduced intelligence, or brain damage. Instead, the research shows that active oversight matters. Higher override behavior was associated with stronger confidence, while heavier reliance on AI output was associated with lower confidence.
The paper argues that AI’s positive effects may be greatest when people use it as a thought partner rather than a substitute for judgment. It also points to future design opportunities: AI systems that prompt users to test assumptions, compare alternatives, explain reasoning, and refine outputs may better support human cognition.
Public discourse around AI often swings between hype and fear; the study offers a more precise message: AI is not inherently harmful or helpful. Its impact depends on whether people remain mentally present while using it.
Source : ID Quotient Advisory Group
