Duke Study Finds AI Use at Work Reduces Perceived Competence, Diligence

A series of studies by Duke University researchers Jessica Reif, Richard Larrick, and Jack Soll reveal that employees using AI tools like ChatGPT, Gemini, or Copilot are often perceived negatively by colleagues and hiring managers, seen as less intelligent, diligent, and more lazy despite AI's productivity benefits. Conducted through four online experiments with 4,400 participants, the research found that this bias diminishes when AI use is appropriate for tasks and when observers themselves have experience with AI. The studies highlight a significant social perception barrier to AI adoption in workplaces, as workers fear being judged as incompetent or easily replaceable. However, firsthand experience with AI increases acceptance of its use. Meanwhile, the evolving tech landscape is influenced by AI's growing role, with companies shifting from a scarcity to an abundance of code, changing employment dynamics in Big Tech. Additionally, frameworks for human-AI collaboration emphasize preserving context and seamless task handoffs to optimize productivity and adaptability in enterprises.