Bridging AI Inequality in Digitally-Mediated Gig Work
Overview
The rise of digitally-mediated gig work, such as ridesharing and food delivery, has led to an increasing prevalence of inequality between platforms and workers. This inequality is exacerbated by thegrowing influence of artificial intelligence (AI), resulting in a critical social issue.Prior literature and our formative studies identified two key factors that drive this AI inequality:(1) the technology divide, where platforms use advanced AI systems that allocate resources, dispatchtasks, and determine pays in the best interest of the platform, while workers lack access to AItechnologies that can support work-related planning and decision-making in their best interests;and (2) the data divide, where platforms have access to comprehensive data from all drivers and customers for planning and operations, while workers only have data about themselves. These two divides lead to unfair working conditions, a growing pay gap, and neglect of workers’ diverse needs. Such inequality also leads to workers’ lack of trust in the platforms.
We propose to address AI inequality in gig work by designing a bottom-up network of intelligentpersonal assistants that empowers gig workers to leverage AI assistance. Each assistant supports a worker’s unique goals, constraints, and contexts, helping them plan for their work and make work-related decisions in their best interests. Each worker can instruct the assistant about their unique goals, constraints, and work contexts through an easy-to-use interface. Assistants will provide recommendations and explain their rationale in a manner that fits into the worker’s existing work context. A network of workers and their assistants will share data, which enables the predictive modeling of task demand and supply, customer and worker behaviors, and pricing changes. This bottom-up approach, which collects data from individual workers, will be compatible with existing gig work platforms (e.g., Uber, Lyft) without requiring additional data access or support from platform operators.
Intellectual Merit
This research will contribute to (1) technical innovations in a system that empowers gig workers to address AI inequality; (2) new algorithmic techniques for modeling AI inequality and stakeholder behaviors using data from an intelligent assistant network;(3) fundamental insights into characterizing and measuring AI inequality in gig work data collection, modeling, and deployment; (4) understanding diverse worker needs and interaction strategies for effective worker-AI partnerships through community-based participatory design and deployment. The contributions span across multiple research disciplines, including Human-Computer Interaction (HCI), Machine Learning, Data Mining, Labor Economics, and Sociology of Labor. Although the scope of this proposed research focuses on empowering workers within the existing paradigm of corporate-owned platforms, its findings will also shed new light on the design of future community-owned gig work platforms.
Broader Impacts
Through our partnership with Chicago Rideshare Advocates, a community labor advocacy organization, the proposed research will (1) make significant real-world impacts by empowering gig workers to optimize their unique goals and interests when planning for work; (2) bridge the AI inequality in gig work through the design and deployment of new technologies. The outcome will specifically benefit underrepresented and disadvantaged workers who are disproportionately affected by AI inequality. The new insights on the current state of AI inequality and worker-AI partnerships can inform broader design and policy implications. Additionally, the research will contribute to the education and training of students, with a focus on involving underrepresented students as the research team have done in the past.