We examine the potential monetization strategies of a video-sharing platform using a game theoretical model, which captures the intricate interactions among the platform, video creators, and viewers. Revenue generated from viewer video consumption is shared with creators via an advertising-based model, a subscription-based model, or a hybrid model of both. Within this ecosystem, the platform balances the video demand and supply, and takes into account the interplay among advertising revenue, subscription revenue, and creator payments. Because of these complex dynamics, the insights from our model require a thorough examination and comparison to understand the equilibrium of the ecosystem under each revenue model. Specifically, we analyze and compare the impacts of each revenue model’s equilibrium on the platform’s profit, content provision, provider earnings, and viewer utilities. We find that the platform owner favors the hybrid model over the pure models when ad nuisance is relatively high but not extreme. Furthermore, we show that the pure subscription model maximizes video quality and creator profits when the production cost is relatively low. Counterintuitively, a higher level of ad nuisance may enhance viewer utilities under the hybrid model, and a higher production cost can bolster creator profits across all three revenue models. Our findings contribute to the literature on video-sharing platforms and offer practical insights to aid platform managers in selecting a revenue strategy.
An increasing number of firms use AI to generate product recommendations. However, when AI-assisted content is delivered to customers, firms face a strategic choice: should the outgoing message be labeled as coming from an AI or from a human? This paper investigates the differential effectiveness of AI-labeled versus human-labeled recommenders across three message framing strategies – personalization-framed, discount-framed, and popularity-framed. Drawing from a field experiment, we find that human-labeled recommenders are more effective in personalization-framed messages, while AI-labeled recommenders are more effective in discount-framed messages. In popularity-framed recommendations, both labels exhibit comparable effectiveness. A follow-up online experiment identifies consumer perceptions of interpersonal and analytical competence as underpinning mechanisms. This research offers theoretical insights and practical implications for businesses aiming to optimize source labeling decisions in AI-assisted recommendation and customer engagement strategies.
The use of virtual influencers (VIs) has emerged as an alternative to human influencers for marketing initiatives. Prior studies have focused primarily on the visual characteristics of VIs and their use in advertising settings, with limited attention given to their capacity to drive pro-environmental behaviors (PEBs). We have developed an integrated persuasive framework, grounded in the Elaboration Likelihood Model and Congruity Theory, that links VI characteristics, message appeals, and behavioral context to persuasion outcomes. Study 1 examined VI message strategies and found that competence (vs. warmth) anthropomorphic conversational style and influencing (vs. de-influencing) appeal enhanced message persuasiveness. Study 2 explored the effectiveness of VIs in promoting plastic waste reduction behaviors across different behavioral contexts and found that messages pertaining to difficult (vs. easy) and private (vs. public) behaviors significantly enhanced message persuasiveness, but that messages promoting easy and private behaviors generated stronger behavioral intentions. The insights from these studies can provide valuable guidance to businesses, marketers, and environmental advocates on effective VI message strategies for addressing challenges related to climate change.
Artificial intelligence (AI) has become increasingly embedded in data-driven online food delivery platforms, significantly improving operational efficiency while reshaping how platform work is organized and governed. However, as AI-based intelligent delivery systems become more deeply integrated into platform operations, increasingly precise scheduling control and strict delivery requirements may raise ethical concerns, including behavioral manipulation and perceived oppression, which are associated with food delivery riders’ job burnout. This paper examines the relationship between AI ethical level and takeout riders’ job burnout, as well as the mediating role of algorithmic trust and the moderating role of organizational trust. Using survey data from 300 takeout riders and the PLS-SEM method, we show that AI ethical level is negatively associated with takeout riders’ job burnout. Furthermore, we identify that algorithmic trust plays a mediating role, while organizational trust has a moderating role. Overall, these findings advance our understanding of the association between AI ethics and job burnout and provide insights into ethical AI design and the trustworthy governance of platform algorithms in data-driven takeout platforms. They also offer practical guidance for takeout platform managers seeking to optimize the ethical design and governance of AI-based intelligent delivery systems and address takeout riders’ job burnout.
Drawing on dynamic capability and resource-based theories, we theorize that team learning is a microfoundation through which firms develop artificial intelligence (AI) capabilities, and that team creativity is a value-realizing mechanism through which these capabilities generate service innovations and improve organizational performance. Study 1 examines how team learning contributes to the development of firms' AI capability and subsequent service innovations, using data from 600 managers in service firms in China. Study 2 investigates whether team creativity mediates the effects of AI capability on service innovation and performance, based on data from 276 managers in service firms in China. Across two studies, we found that teams contribute to the development and realization of AI capability. Specifically, team learning facilitates the development of AI capability, and team creativity enables firms to translate this capability into innovation and performance outcomes. Overall, this study advances the understanding of how human–AI collaboration enables firms to convert AI capability into tangible organizational value by integrating a microfoundational perspective with the resource-based logic of value realization.
