“Capacities for social interactions are just being absorbed by the model”: User engagement and assetization of data in the artificial sociality enterprise

Developed by a Korean start-up called ScatterLab, Luda Lee is an AI-powered social chatbot designed to embody the persona of a 20-year-old, youthful college girl and promoted as “your first AI friend.” A year after the suspension of Luda 1.0 in 2021 amid a heated controversy surrounding Luda 1.0 in 2021 regarding user-initiated abuse and the unethical sourcing of language data, ScatterLab recently reintroduced a generative AI-powered Luda 2.0 to the public, while pursuing the expansion of its operation as a software-as-a-service provider. The article traces the evolution of both Luda and ScatterLab, highlighting the crucial role of user engagement in the making and repurposing of the company’s “proprietary dataset.” It explores how user engagement is not only an objective but also a strategy for corporate growth in the artificial sociality enterprise. Strategies to encourage user engagement, like gendering of the chatbot design, create unanticipated kinds of user engagements and user-generated data. Those data, conceived as materials from which the AI’s competence is absorbed, is constructed as an asset, a material and discursive anchor for the company’s future promises.