AI agents are not your “coworkers”

MIT Tech Review's editorial argues that marketing AI agents as 'coworkers' or 'digital employees' is misleading and potentially harmful. The article points out that anthropomorphizing AI tools blurs the line between human workers and software, leading to unrealistic expectations about capabilities, accountability, and trust. Instead of treating AI as a peer, the piece suggests developers and managers should view agents as tools with specific strengths and limits—akin to advanced calculators, not colleagues. For builders integrating AI agents into workflows, this distinction is crucial: it affects how we design human-AI collaboration, set user expectations, and handle errors. The author warns that overpersonalization can obscure the need for human oversight and lead to over-reliance on unreliable outputs. The practical takeaway is to name and frame AI tools transparently, avoiding language that implies agency or responsibility. This aligns with a growing consensus in AI safety that clear role definitions reduce misuse.