Vernacular AI: the bottom-up domestication of innovation

Led by: Soumodip Sarkar

This keynote argues that the most consequential AI shift is not top-down strategy but vernacular AI: the bottom-up domestication of general-purpose models into local entrepreneurial routines through prompts, workarounds, and platform features. Yiwu merchants are an instructive case of this dynamic, using pragmatic digital bricolage to survive and sell across borders rather than merely to optimize. As access to generative capability lowers the cost of producing content, offers, and outreach, the challenge shifts from creation to differentiation and trust: when many actors draw on similar models in similar ways, novelty becomes easier to generate but harder to defend, evaluate, and verify, creating a “sea of sameness.” To explore this, I propose four archetypal use-patterns for businesses using AI (and a parallel coding frame for researchers): situations where AI removes friction in routine operations; where it scales the volume of market experiments; induces convergence and displaces judgment or accountability; and where value depends on provenance, verification, and stakeholder trust. The talk concludes with a research agenda that treats AI less as a standalone firm capability and more as ecosystem infrastructure, urging scholars to measure when these tools broaden opportunity for SMEs versus when they commoditize novelty.