Economically, the distinction between automating jobs and augmenting tasks is pivotal. Research warns that “so-so automation” can replace workers without large productivity gains, dampening wages (Daron Acemoglu, 2020). By contrast, complementarity—tools that raise human productivity—can expand demand for expertise (David Autor, 2015).
Consequently, design choices are distributional choices. If AI amplifies clinicians, teachers, and tradespeople, value accrues to workers and customers, not just platforms. Policies can reinforce this direction: incentives for augmentation-focused R&D, training subsidies, and procurement standards that reward human productivity and safety, not headcount reduction alone. [...]
