Skill-biased technological change complements workers whose knowledge helps them use new tools while substituting for selected routine activities. The technology can increase output without reducing total employment if demand expands or workers shift toward tasks that remain valuable.
AI may favor judgment, context, evaluation and the ability to reorganize work around model capabilities. The distributional effect depends on access, training, adoption and whether productivity gains create new products and demand.
ELI5
Skill-biased technological change makes some human skills more valuable or productive than others. New tools may replace selected routine activities while increasing demand for people who can direct, judge, combine, or improve the technology.
For example, AI may automate a basic draft but increase the value of domain judgment, verification, and workflow design. The effect on jobs and pay depends on access, training, adoption, and whether productivity creates new products and demand.
How can AI be skill-biased?
It can make workers with strong domain judgment and tool-use skills more productive while automating routine activities performed by other roles.
Does skill-biased change always reduce employment?
No. It changes relative productivity and demand, while total employment also depends on growth, new work, wages and how organizations adopt the technology.

