Artificial Intelligence (AI) users who let robots do their thinking at work are taking the easy route, which is harmful to their minds, a new study has found. A survey of 300 employees at Microsoft, the maker of the Copilot AI system, found “self-reported reductions in cognitive effort” among many users who use AI at work in more than 900 ways. If used “incorrectly, the technologies result in a reduction in cognitive skills that need to be preserved,” the scientists warn in a paper to be presented in late April and early May at a conference on computers and human factors in Yokohama, Japan.
“Higher trust in artificial intelligence is associated with reduced critical thinking, while higher self-confidence is associated with stronger critical thinking,” claims a team of scientists from Carnegie Mellon University.
In other words, the use of artificial intelligence could result in humans being left with mundane, simple tasks like controlling the often unpredictable behavior of robots, rather than the more difficult cognitive tasks of creating strategies and solving problems, the research suggests. “Generative artificial intelligence shifts critical thinking toward validating information, integrating responses, and managing tasks,” the Microsoft-CMU team reported after assessing the responses of its employees who used robots like ChatGPT for tasks like writing emails or reviewing their colleagues’ work.
