Artificial intelligence (AI) chatbots and related technologies are increasingly implemented in organizations worldwide to improve efficiency and accelerate employees’ and teams’ productivity. This exploratory and conceptual paper examines some unfortunate and unin-tended consequences for computer-supported cooperative work when AI systems for the workplace are designed in anthropomorphic ways. Our, perhaps provocative, argument is that emerging collaborative technologies in the workplace should be designed to appear less, rather than more, trustworthy. This is because AI systems (1) are characterized by an-thropomorphism, (2) employ rhetorical conversational strategies that persuade users through knowledge claims, and (3) inherently represent uncertainty, as these technologies possess wide-ranging capabilities and evolve rapidly. Together, these characteristics may increase the risk of unfortunate decisions by knowledge professionals. We conclude by em-phasizing that AI design for the workplace, the use of AI by knowledge professionals, and the epistemic power of workplace AI systems are important topics that the CSCW research com-munity should address and place more prominently on its research agenda.
Pettersen et al. (Thu,) studied this question.