Research
Work in the United States is moving away from long-term employment toward contingent, project-based arrangements increasingly mediated by digital systems. Three shifts organize my research: the growth of digitally enabled remote work, the increasing centrality of knowledge and service work in the economy, and a change in the unit of employment from the career to the job to the project. Online labor platforms lie at the intersection of these shifts; I study them because work there is already found, priced, and performed through digital systems.
I examine this transformation at three levels — the individual worker, the organization, and the technology — using longitudinal qualitative methods together with human-centered design. My training is in the sociology of work, and my research also draws on management, labor studies, information science, and human-computer interaction.
Using and competing with AI-enabled systems
The most recent stream of my research examines what happens when AI-enabled systems enter work that is already digitally mediated. Online freelancers encounter these systems on two sides simultaneously: they adopt tools such as ChatGPT, Copilot, and Claude to perform their work, and they compete with the same class of systems, because they are hired for the discrete, knowledge-based tasks those systems are designed to perform. A panel study conducted with Steve Sawyer (Syracuse University), building on an established longitudinal panel, asks how these workers use AI-enabled systems, how they compete with them, and how workers and systems change over time.
Early findings indicate that the effects are neither simple substitution nor simple assistance. Some workers adopt these tools to raise their productivity and competitiveness, while clients, assuming that the work is now AI-assisted, seek corresponding reductions in pay; the value workers create through their own use of the technology is thus partly appropriated by the firms that hire them.
AI-enabled tools
Online labor markets are designed in ways that favor the buyers of labor: platform policies tend to advantage employers, and the systems that match workers to jobs and set prices are opaque to the workers subject to them. A related stream asks whether the same class of technology can instead be built to serve workers. Conducted with online workers and the organizations that support them, it designs, deploys, and evaluates digital tools — including AI-enabled ones — that help workers represent their skills, identify well-paying and skill-building work, and secure recognition for the quality of their work.
The work uses human-centered and value-sensitive design: interviews and focus groups with online workers establish the functionality they require, and prototypes are refined across successive cycles in collaboration with worker-serving organizations. Its premise is that this class of technology is not inherently employer-serving, and its aim is to produce both working systems and evidence about which capabilities workers most value when they specify the tools themselves.
Transactional careers: how digitally mediated work reshapes working lives
My collaborators and I have followed a panel of online knowledge workers through more than 450 interviews, examining how careers, identities, and livelihoods take shape when work is bought and sold through digital platforms and, because the study is longitudinal, how those effects accumulate over time rather than appearing in a single cross-section. The study treats platforms not as neutral marketplaces but as market-makers whose architecture and interfaces mediate the relationship between workers and the firms that hire them.
From this work we have developed the concept of the transactional career: a working life organized not as a continuous employment history but as a sequence of discrete projects, valued for its flexibility and constrained by a persistent need for stable work. We show how platforms decompose professional identity into platform-specific metrics, how the resulting platform-specific capital does not transfer to settings off the platform, and how these processes reproduce, and in some cases attenuate, established inequalities of gender and race. The analytical strategy combines cross-sectional comparison within waves, longitudinal comparison across waves, and comparative case analysis, with structured case studies of each respondent supporting fuzzy-set qualitative comparative analysis.
Digital entrepreneurship
A further stream examines the organizations that are founded inside platform infrastructure. I study founders who built agencies within Upwork, converting individual freelance work into firms that recruit, coordinate, and sell the work of others. The setting is analytically distinctive because the market, the rules, and much of the operative technology belong not to the founder but to the platform.
The study examines the attributes of these founders and the organizational structures they create, and it reconsiders two established constructs — entrepreneurial orientation and entrepreneurial skill — for a context in which entrepreneurs build atop infrastructure they do not control. It is designed as a pilot for a longitudinal study of platform-native firms that would parallel the worker panel.
About the digital entrepreneurship project →
Flexibility, beyond scheduling
Flexibility is among the defining features of non-standard work and the common element across its many forms — compressed hours, flextime, co-working, contracting, platform work, micro-tasking, nomadic work, and portfolio careers. Scholarship on platform workers has concentrated almost entirely on one dimension of it, flexible scheduling, in part because scheduling flexibility is structurally intrinsic to contract work.
Comparatively little attention has been paid to the other conditions that create or erode flexibility for platform workers. My work distinguishes those dimensions, among them task and spatial flexibility, and shows how gender, occupation, and periods of crisis — the pandemic in particular — shape how much that flexibility is worth to workers.