Complementor Value Co-Creation in Generative AI Platform Ecosystems
研究了生成式AI平台中互补者如何通过四种机制（系统指令、上下文数据、用户输入筛选、模型输出修正）应对开放性和不可解释性挑战，实现价值共创。
Compared to traditional digital technologies, generative artificial intelligence (GenAI) exhibits two facets—open-endedness and inscrutability—that create challenges for the value co-creation of complementors in GenAI platform ecosystems. For example, complementors lack standardized, reusable functionality that powers their applications and struggle to consistently generate desired outputs. To identify how complementors mitigate these challenges, we conducted an embedded case study of complementors in the OpenAI platform ecosystem. We identified four value co-creation mechanisms of embedded and stand-alone complements that mitigate the challenges: (1) utilizing system instructions, (2) providing context data, (3) curating user inputs, and (4) revising AI model outputs. Depending on the complement type, complementors employ these mechanisms, leverage synergies between mechanisms, and unlock previously not employable mechanisms. We contribute to research by revealing two novel logics—the reap logic and the differentiation logic—that explain how complementors co-create value by mitigating the challenges through the value co-creation mechanisms.