The future of the labor force: higher cognition and more skills

Skills can be divided into social-cognitive and sensory-physical. Sensory-physical skills follow explicit rules and can be executed via programming, leading humans to trust machines with such tasks. Social-cognitive skills involve open interpretation, context dependence, and intuition, making them harder to codify and less suitable for machine replication due to Polanyi’s paradox and algorithm aversion. Machines are specialized and costly to retask, while humans switch tasks more readily; thus, versatile workers are less automatable. Using O*NET skill scores and OEWS employment data for the U.S., the study empirically validates that: higher cognitive skill scores reduce automation susceptibility; workers with a broader array of skills gain employment share; and jobs emphasizing sensory-physical skills are more likely to be replaced. Two strategies emerge for workers: build diverse cognitive skills (e.g., creativity, critical thinking) and build diverse overall skill portfolios (social-cognitive plus sensory-physical). Specialization in a narrow sensory-physical skill offers little advantage. The study provides the first empirical evidence of differential automation impacts on these two skill sets and delineates traits of irreplaceable labor, informing career development and education policy.

Wen Zhang, Kee-Hung Lai, Qiguo Gong