AI Agents Should Not Be Considered Coworkers

Picture arriving at your office to find that an AI tool named Alex has been designated as your new subordinate. This scenario raises questions about how effectively you would collaborate with Alex. Research led by Emma Wiles from Boston University indicates that when managers view AI as a coworker rather than a mere software tool, their performance declines. Specifically, they identified 18% fewer errors when the AI was labeled as an 'employee.' This finding highlights the significant impact of terminology on workplace dynamics. The trend of marketing AI as digital colleagues is gaining traction, with major tech companies promoting these tools as capable of functioning like humans. However, this perception can lead to unrealistic expectations and diminished accountability among human workers. Wiles's study found that managers were less likely to take responsibility for AI-generated outputs when they were framed as employee contributions. This shift in perspective could have serious implications across various sectors, including healthcare and education, where AI may be unfairly blamed for failures stemming from human oversight. Experts argue that AI should enhance human capabilities rather than replace them. A Stanford initiative demonstrated that while workers see value in automation, they often disagree with tech experts on which tasks should be delegated to AI. Ultimately, labeling AI as employees may simplify accountability but does not improve its effectiveness, leaving human workers at a disadvantage.