Organizing Intelligence
A powerful AI model does not automatically make a powerful organization. In Organizing Intelligence, Alex Imas, Andrew Koh and Shane Legg ask how human and artificial capabilities combine inside firms, and why coordination may matter as much as individual expertise. Their theoretical model separates improvements in task competence from improvements in working together. That distinction changes the possible paths of firm size, market concentration, wages and employment. Complementary human skills can sustain productive teams even when AI outperforms the average individual; stronger AI coordination can, under specified conditions, reorganize production around much larger firms. The paper also distinguishes gradual replacement within mixed teams from abrupt displacement when AI firms outcompete human firms. These are conditional economic mechanisms, illustrated by mathematical examples rather than forecasts of when jobs will disappear. The study offers a demanding question for AI evaluation: how should we measure the capabilities of an organization, rather than those of a model working alone?
