UCLA

As technology continues to advance rapidly, understanding its broader societal implications becomes increasingly important. This thesis explores the intersection between human and emerging technologies—specifically generative artificial intelligence (AI), autonomous vehicles (AVs), and hybrid marketplaces—and their potential impacts on society. Through three interconnected studies, we investigate how these technologies influence user behavior, market dynamics, and service quality, providing valuable insights for policymakers.

In Chapter 2, we delve into the interaction between humans and generative AI. While generative AI can boost productivity, the content it produces may not always align with user preferences. To study this effect, we introduce a Bayesian framework where heterogeneous users decide how much information to share with the AI, balancing a trade-off between output fidelity and communication cost. We reveal that these interactions can lead to societal challenges such as homogenization and bias. Our findings highlight the risk of reduced diversity in outputs, especially when AI-generated content is used to train the next generation of AI systems, potentially resulting in a “homogenization death spiral.” We also assess the impact of AI bias, demonstrating how AI biases can lead to societal bias. Importantly, we suggest that facilitating human-AI interactions can mitigate these risks.

Chapter 3 investigates the societal effects of introducing AVs into a ride-hailing market which is currently served by human drivers (HVs). We develop a game-theoretical queueing model in which a platform aims to maximize its profit while HVs make strategic joining decisions. Our analysis indicates that incorporating AVs may degrade service levels, as the platform may prioritize AVs, negatively impacting HVs’ earnings and driving them out of the market. We then reveal that this reduction in service level is not uniform in a city: high-demand areas, such as downtown areas, may maintain reasonable service levels, while remote areas may experience a large decline in service level. Then, using New York City data, we build a highly detailed simulation of the operations of a ride-hailing platform to further validate our theoretical model in a more realistic setting and demonstrate the additional effects on service levels. This study underscores the importance of balancing profitability with service quality when introducing AVs in the transportation sector.

In Chapter 4, we extend the analysis of Chapter 3 and focus on the strategic decisions of profit-maximizing firms operating in “hybrid marketplaces” consisting of both private and flexible supply agents. The firm can decide the number of private agents to employ, paying them regardless of their work, while flexible agents make their own revenue and pay a commission to the firm. We develop a general framework for supply prioritization, applicable to any firm using a mix of employees (private agents) and contractors (flexible agents), and capable of handling complex supply management policies. Our findings show that without prioritization, using hybrid supply is not optimal. However, effective prioritization strategies can enhance profitability by increasing the productivity of private agents, albeit at the cost of reducing flexible supply participation. Therefore, the firm tends to prioritize private supply in “over-supplied” markets, but may prioritize flexible supply in “under-supplied” markets, where a slight increase of supply increases can significantly impact outcomes. These insights highlight the critical role of prioritization in managing hybrid supply in markets.