UNIVERSITY PARK, Pa. — The latest McKinsey Global Study on the State of AI found that approximately 88% of organizations worldwide will have implemented artificial intelligence (AI) in at least one business function by the end of 2025. Despite promising productivity gains, passive use of AI in the workplace, where employees complete tasks by copying and pasting the AI’s responses, can make people question their own skills and feel less meaningful at work, according to a study published in Scientific Reports and co-authored by faculty at Penn State’s Smeal College of Business.
Using Prolific, an internet platform designed to help scientists find research participants, the team recruited approximately 270 professionals working in human resources, communications, and management fields to complete a series of written tests similar to their daily work, both manually and with the help of AI tools. Their study found that the use of AI, specifically whether participants used the AI collaboratively to workshop their ideas or passively to generate and copy responses, played an important role in participants’ reported scores of self-efficacy, meaningfulness, and psychological ownership. Specifically, the researchers found that passive AI use reduced feelings of ownership by nearly 20% and perceived meaning by 10%, while collaborative AI use showed similar scores to tasks that did not rely on AI.
Although the use of AI has been reported to improve productivity, less is known about the deeper psychological effects of using AI in the workplace, explained Yidan Yin, assistant professor of management and organizations at Penn State’s Smeal College of Business. Professor Yin explained that although scientists are beginning to investigate the potential long-term costs, the field is still very new and much of the research is very wide-ranging.
“Previous studies have mainly considered the positive impact of AI on work productivity and the potential for its use to isolate and demotivate workers,” Yin said. “But in this study, we wanted to focus on better understanding how the use of AI reshapes the way people connect to work.”
To accomplish this, the team primarily focused on measuring the impact of using AI on three closely related concepts. It is self-efficacy, or the confidence of an individual to complete a task without the assistance of an AI. Meaningfulness of work, or the extent to which individuals perceive their work as purposeful and important. and psychological ownership, or the extent to which individuals feel ownership of their work product. Yin explained that the researchers used two additional variables: task enjoyment and outcome satisfaction to get a comprehensive picture of how the use of AI affected participants’ psychology.
The researchers constructed a series of writing tasks tailored to the occupations of the study participants. For the first task, participants were assigned to one of three conditions and instructed to either complete the task manually without the AI, actively work with the AI, or passively copy and paste the AI-generated responses to complete the task. Participants then answered questions about self-efficacy, meaning in their work, and psychological ownership of their accomplishments. The second task required all participants to manually complete a writing task without AI assistance and then answer the same survey questions.
“This two-task design allowed us to examine both the immediate effects of different uses of AI and the lasting effects after participants returned to work without AI in an experiment that took only 20 to 30 minutes to complete,” Yin said.
Passive AI use in the first task reduced people’s sense of ownership by nearly 20% and self-efficacy and perceived meaning by nearly 10% compared to handwriting, whereas collaborative AI use was not significantly different from manual creation. The decline in self-efficacy and meaningfulness persisted after the second task, when all participants returned to handwriting, suggesting that the decline cannot be easily reversed by returning to work without AI assistance.
