Abstract
Our paper formalises a production function to give microeconomic foundations for the adoption of Generative AI (GAI) within workplaces, by modelling the process of task allocation to workers and GAI in a context of tasks-interdependencies. We assume that human-created and AI-generated outputs cannot be distinguished. We formalise non-market interactions between workers and technology to characterise closed and open learning processes. We demonstrate that the skills to detect errors are local, collective and technologically interactive. Finally, we analyse several organisational configurations linked to the adoption of GAI and the proportion of tasks allocated to workers. Our results shows that GAI should be matched with most able workers.

This publication has 3 references indexed in Scilit: