The emergence of platform-based work over the past decade disrupted labor markets across the globe. As of Sept 2023, the gig workforce was estimated to range from 154 to 435 million workers, comprising 4-13% of the global labor forcefoot_0 [13]. Workers increasingly engage in platform-based gig work for the promise of work exibility and autonomy [40], potential to mitigate discrimination as enabled by anonymity on certain platforms [20] and opportunity for upskilling provided by macrotask/freelancing platforms [20].
But as platform-based labor emerges to complement traditional employment, workers face unprecedented challenges and data harms [34]: algorithmically-reinforced inequality and power differentials [6,7,26,40], overexposure to workplace monitoring and surveillance [29,30], physical risks [2,11,32], heightened uncertainty [3,28], and social isolation [41,42]. Numerous nations intend to increase regulation of labor platforms [10,14,37], but are limited by the scarcity of publicly accessible worker data [21].
In resistance to surveillance and hegemonic data practices of platforms [1,33,35], workers increasingly engage in self-tracking through individual means [22] or third-party tools 2 . In the absence of sucient policy and regulations for responsible platform practices, researchers and advocates increasingly turn to data collectives and tools as a method for advancing regulation [7,25], restoring worker power [17,24,36,43] and holding platforms accountable to more ethical, fair and community-centered data practices 3 [29].
To dene worker data collectives, we turn the HCI/CSCW literature for aggregating potential future data infrastructures [24,36,44]. Recent eorts leveraged participatory design with workers and relevant stakeholders to reveal several (counter-)data collectives for supporting workers. Such collective data institutions included digital social institutions (e.g., collective wikis, online forums/groups/unions [42]), oine social institutions (e.g., union strikes leveraging social media to coalesce/organize [23]), thirdparty tools [24,36,43], self-tracking [22], and platform-evaluation (e.g., Fairwork [19]). Regardless of the specic infrastructure, data collectives hold considerable promise for facilitating worker advocacy and empowerment, since they embody a site for communities of resistance [4] and enable collective data actions (e.g., counterdata collection, data refusal/strikes [38,39,45]).
To fully enact the potential of data collectives as a vehicle for producing counter-data and restituting worker power/rights, designers and maintainers must prioritize principles of care [5,15,17], ethics [27] and justice [12,18]. We draw from seven principles of the intersectional feminist framework by D'Ignazio and Klein [16] and insights around workers' challenges informed by prior empirical work [24,25] to consider ways of:
Articulating invisible/unpaid work and addressing wage theft-Principle 7: Making Labor Visible.
Collectively auditing/disaggregating worker data withheld by platforms and challenging resultant algorithmic decisions-Principles 1 & 2: Examining & Challenging Power.
Addressing (physical and digital) safety risks that platforms fail to account for, including dangers present on roads, in strangers' homes, and from online scams-Principles 3 & 6: Elevating Emotion and Embodiment by Considering Context.
Gathering qualitative accounts/narratives of discrimination against marginalized individuals and work strategies-Principles 4 & 5: Rethink Binaries and Hierarchies, Embrace Pluralism.
Building infrastructure around interpreting and operationalizing assets in data collectives to precipitate material change-Principle 6: Considering Context.
Ultimately, advocates leveraging data collectives aim to improve labor regulations or propose litigation to advance worker (data) protections. To ensure policy-inuencing data collectives maintain long-term trust with workers, designers must consider the balance of governance/power structures with privacy protections, while allowing non-worker stakeholders to access necessary insights to make informed decisions. In light of such multi-stakeholder considerations, we plan to discuss eective designs to unlock potentials of data collectives as boundary objects to connect dierent stakeholders' needs and ways of knowing and collaborating, where stakeholders include 1) workers, 2) researchers, and 3) practitioners (advocates, activist groups, lawmakers and policymakers, etc.).
3 e.g., FairFare, a worker auditing tool to uncover platform commission, and Driver's Seat Cooperative, now under the Worker's Algorithm Observatory, to help drivers and researchers investigate gig platform transparency and workers' experiences
Convene a community of dierent stakeholder groups to discuss challenges and opportunities of worker data-sharing collectives for empowering platform workers. Many researcher, advocacy, and worker-organizing eorts have converged on the importance and necessity of worker data (practices) for auditing platforms, surfacing platform manipulation, or informing the need for policy and regulation [8,9,31,36,43]. This workshop will serve as an avenue for collaboration among these existing eorts. Contextualize worker data within broader questions of worker rights, well-being and autonomy, including asking what kinds of worker data are meaningful, where data is shaped by conditions of constant worker surveillance, and the limitations of data as a tool. Ideate and exchange perspectives on how such technologies can be governed and impact labor regulation across geographic regions/nations. In addition to constructing a shared understanding of the landscape, we aim to form a future research agenda.
A tentative workshop schedule is outlined in Table 1. We will begin with a welcome keynote by 1-2 speaker(s) experienced in worker advocacy or labor policy. Next, participants will introduce their backgrounds and interests through lightning talks. Following a break, participants will engage in interactive group design and discussion to document ideas, themes, experiences, challenges/questions, and resources related to worker data collectives. Afterwards, each group will present the outcomes of their design. The workshop will conclude with a synthesis of high-level themes surfaced from presentations and a discussion of future directions.
To begin the workshop, 1-2 keynote speaker(s) with rsthand experience at/with (non-prot) worker organizations will share insights on challenges and opportunities related to labor advocacy for platform workers. We will extend the invitation to active workerorganizations (e.g., Rideshare Drivers United and Colorado Independent Drivers Union), non-prot institutions (e.g., Colorado Fiscal Institute), legal advocacy groups (e.g., Towards Justice), and leading academic researchers.
Worker Data Collectives as a means to Improve Accountability, Combat Surveillance and Reduce Inequalities CSCW Companion '24, November 9-13, 2024, San Jose, Costa Rica
Participants will introduce themselves and share reections on a question below as addressed in their submissions. Listening participants will be encouraged to respond with further reections.
Participants will brainstorm potential challenges in designing a worker data collective and issues of current platform work conditions for the system to address. Below are higher-level questions around data collective design and an overview of potential stakeholders, issues and data structures (Fig. 1) to kick-start the session. Activity 3: Co-Designing Worker Data Collectives
Participants will break into groups. Each group will design data collective structure(s) for a specic platform/work type using digital templates (e.g., guided Miro boards) and/or physical materials (e.g., posters, sticky notes, markers). Participants' submissions will inform their group assignments. Examples of possible platform groupings include: 1) Rideshare & Delivery (e.g., Uber, Doordash), 2) Freelancing & Macrotasking (e.g., Upwork, Fiverr), 3) Microtasking (e.g., Amazon Mechanical Turk, Crowdower, Appen), 4) Caretaking and Household Work (e.g., Care.com, CareRev). When designing worker data collectives, we encourage participants to consider the following questions:
Each group will present their data collective from activity 3. This can include describing infrastructural decisions, ideas for addressing the design questions, and new concerns or questions that arose during discussions. Observing groups will be encouraged to ask follow-up questions and share reections, while keeping in mind the questions below:
To frame the nal discussion, facilitators will summarize opportunities and challenges based on participants' ideas, questions, and concerns. Participants will be given space to consider and propose future research agendas or avenues of work.
Post-workshop, a document will be shared to participants to summarize each group's designed data collective with a) a link to the correlating Miro board, b) photos of physical artefacts created if applicable, c) a summary of the group's presentation and questions surfaced by others, and d) questions and themes from the talk-back session. Furthermore, we seek to support continuing collaboration interests that arise-for example, we may create a shared document for participants to share new resources or set up a collaborative platform to facilitate cross-organizational eorts related advancing work data collectives. Inspired by the workshop by Yang et. al. on bridging HCI and policy design, we may also consider synthesizing workshop insights into a provocation/position paper.
This workshop will run as a full-day hybrid workshop to allow participation from a diverse range of geographic locations and backgrounds. Sessions will be mediated through Zoom and asynchronous conversations will be facilitated via Slack .
We will recruit a maximum of 50 participants who work on or demonstrate interest in platform-based labor. This includes researchers with backgrounds in Computer-Supported Cooperative Work, Human-Computer Interaction, Public Policy, Law (and beyond), as well as organizers, activists, and platform workers.
Interested participants should submit a statement of interest as 1) a maximum 500-word personal statement or 2) a maximum two-page extended abstract/case study about a specic type of platform-based work as related to the workshop themes. The statement should address the question: How data can inform policymaking? To optimize group assignments, we recommend submissions specify the type(s) of platforms/work where they have the most interest/experience.
We highly encourage submissions to reect on concepts of power, ethics and their own positionality as related to platform-based work and counter-data. Guiding questions of Activity 1: Lightning Introductions & Reections can provide a starting ground. Submissions incorporating gures/diagrams for ideating data sharing structures are welcomed but not required; gures, diagrams, and references do not count towards the page limit.
Equipment and Supplies Needed to Run the Workshop: To accommodate in-person participants, we request access to standard conference room facilities, including seating for up to 25 participants, A/V equipment, and access to physical design resources (e.g., markers, sticky-notes, posters/whiteboards/large easel pads). Resources participants are expected to bring or provide: Online participants will need access to a desktop computer or laptop with internet connectivity to participate. In-person participants will also be expected to bring laptops in order to participate in the Miro board activities, and optionally to access their own and/or other participants statements of interest.
Lower bound of 154 million or 4.4% represents an estimate of only main/full-time workers while upper bound of 435 million or 12.5% also includes part-time/secondary workers
e.g. Gridwise, Stride and Strava