Emma Wiles, a Boston University business professor, discovered that managers who treated an AI tool called "Alex" as a coworker rather than software caught 18% fewer errors. Her research reveals that framing matters enormously in how humans interact with artificial intelligence in workplace settings.

Microsoft, OpenAI, Anthropic, and Google have all released tools for managing teams of AI agents since April, many marketed explicitly as digital colleagues. Nearly one-third of 1,261 surveyed managers said their companies already frame AI agents as employees, with 23% actually listing them on organizational charts.

When AI tools were framed as employees, study participants felt less responsible for their output and were 44% more likely to escalate questionable work rather than correct it themselves. This responsibility inversion undermines the time-saving purpose of using AI agents and risks creating convenient scapegoats for human failures in critical sectors.

Stanford researchers presented 1,500 workers across 104 jobs with AI task information, finding that workers wanted automation in some areas but rejected others that technologists deemed suitable. MIT economist Daron Acemoglu argues AI should enhance human capabilities rather than replace humans, suggesting current agentic AI development has lost sight of this fundamental purpose.