BOSTON — A recent study led by Boston University business professor Emma Wiles found that managers detected 18% fewer errors when artificial intelligence was presented as an "AI employee" compared to a chatbot. The study, which included 1,261 managers, indicated that framing AI in this manner influenced managers' perceptions of responsibility and their approach to error detection.

Participants in Wiles's study reported feeling less accountable for AI output when the AI tool was characterized as an employee. Additionally, managers were 44% more inclined to refer questionable AI work to another manager for review if the AI was framed as an employee.

The study also examined current corporate practices regarding AI. Almost one-third of the participating managers stated that their companies currently present AI agents as employees. Furthermore, 23% of the managers indicated that their companies include AI agents on organizational charts.

Daron Acemoglu, an economist at MIT, discussed the broader implications of how AI is positioned in the workplace. "AI agents right now are being marketed as things that can replace humans, and I think that's just a losing proposition," Acemoglu said. He added, "They should instead be optimized so that they can improve human capabilities, which is not what they have [been] at the moment."

Why It Matters

This study demonstrates how the language used to describe AI can impact managerial behavior, particularly regarding responsibility and oversight. The findings suggest that presenting AI as an "employee" may inadvertently reduce human vigilance over AI-generated work, potentially increasing the risk of unaddressed errors.