Your AI bot isn’t an employee
AI agents are getting names, friendly faces and, in some companies, a place on the org chart. With tools like Muse, Dots and Grok Bot attracting attention, a Boston Consulting Group study raises a useful question: does treating software like a colleague make us less careful with its work?
The setup: BCG researchers led by Matthew Kropp surveyed more than 1,200 HR and finance professionals in the US, Canada and EU. Nearly a third of managers framed AI as a teammate or employee. More than 20% had put agents on a formal org chart.
The experiment: Participants reviewed the same error-filled document, attributed to a human employee, to "an AI tool," or to a named AI "employee." The named-employee group caught fewer errors, and assigned more responsibility for errors to the AI and less to themselves..
- Those managers also reported more concern about being replaced and less trust in how AI would be used.
Why it matters: Managers reviewed the same document differently depending on who they thought had produced it. Calling AI an employee made them less likely to catch mistakes and more likely to hold the software responsible.
Our take: Call your bot whatever you like. Just make sure the person using it knows they’re still responsible for checking its work.
