Between the Lines: Understanding Employee Preferences for AI Usage Guidelines Through Discrete Choice Experiments

As organizations increasingly adopt artificial intelligence (AI), employees’ primary interaction with AI governance occurs through internal AI usage guidelines. However, little is known about how employees evaluate different guideline features. This study examines employee preferences using a discrete choice experiment. Six guideline attributes were included: error accountability, usage flexibility, policy standardization, AI training, employee feedback, and AI-related job displacement. A Bayesian D-optimal partial profile design generated paired guideline scenarios, and responses from 315 employed participants were analyzed using a panel mixed logit model. AI-related job displacement was the most influential attribute, with replacement-oriented policies strongly disliked and support-oriented policies most preferred. Employee feedback mechanisms and AI training were also highly valued. Participants preferred shared organizational accountability over individual responsibility and evaluated mandatory AI use negatively. Policy standardization had the weakest effect. These findings provide quantitative evidence on how employees evaluate and trade off key features of corporate AI usage guidelines.