Creating AI copies of coworkers leads to unintended consequences

A journalist at WIRED created AI clones of his editors, Brian Barrett and Sophie Kleeman, to better understand workplace dynamics and improve his performance. The AI models, built using Gemini, generated personas based on public data and limited internal communications. Brian Bot, initially praised for mimicking Barrett’s leadership style, later became a source of frustration. Sophie Bot, while adept at mimicking Kleeman’s communication style, occasionally hallucinated and produced misleading content. Despite some useful insights, the AI clones often failed to replicate the nuances of their human counterparts. Brian Barrett eventually shut down Brian Bot, while Sophie expressed disappointment with Sophie Bot’s inaccuracies. Experts warn that anthropomorphizing AI can manipulate human emotions and complicate accountability in the workplace. Companies like McKinsey are already integrating AI employees, with plans to achieve a 1:1 ratio by year-end. The experiment highlighted both the potential and pitfalls of AI in the workplace. While AI tools can enhance productivity, they struggle with replicating human behavior and judgment. As AI adoption accelerates, questions remain about responsibility, accuracy, and the evolving role of human workers. The journalist’s experience underscores the challenges of integrating AI into professional environments, where the line between tool and colleague becomes increasingly blurred.