Faster Isn’t Healthier: How AI Is Reshaping Digital Marketplaces

Online marketplaces run on short-lived product windows: sellers launch, advertise, learn, and exit as opportunities open and close. Generative AI now shortens the time it takes to list a product and pivot to the next one. A new research outcome from the Asia-Pacific Social Innovation Research Institute (APSIRI), authored by Mainak Mondal of the Courant Institute of Mathematical Sciences at New York University, asks a deceptively simple question: when everyone can move faster, does the marketplace as a whole become healthier, or does faster churn deplete opportunities for all?

Using an agent-based model in which AI affects only the cost of listing and the time needed to migrate between products, the study finds that the answer is not obvious in advance. For an isolated seller, faster relaunching can extend survival. But when many sellers adopt AI at once, they crowd into the same visible opportunities, raising saturation and shortening the very product windows they are all trying to exploit. Individual speed does not add up to a healthier market.

The most striking result is the possibility of what the paper calls “zombie markets.” When new opportunities are not replenished quickly enough, a marketplace can remain financially solvent while its active products, seller participation, and assortment diversity quietly collapse. Crucially, metrics that track only trust or seller survival would miss this hollowing-out entirely. The model shows that only a narrow band of moderate platform governance keeps trust, participation, and diversity positive at the same time.

The broader lesson is about how we judge the health of AI-shaped platforms. Rather than treating generative AI as a simple productivity boost, the study reframes it as an auditable reduction in friction whose collective effects depend on replenishment and governance. Platform health, the author argues, must be assessed through activity and diversity together with trust — not through seller survival or trust alone.

This research reflects APSIRI’s interest in computational social science and the governance of digital economies — using rigorous modelling to understand how new technologies reshape markets, and what it takes to keep those markets open, diverse, and trustworthy.