How to Bargain With a Black Box: Hacking a Path to Data-Driven Organizing

Dan Calacci, a PhD student at MIT, argues that the labor movement needs people with technical skills to build tools for workers. He opens with a retailer's delivery platform that moved drivers from a transparent per-order percentage to a black-box pay algorithm in 2020. Drivers began sharing pay screenshots online, and the pattern suggested that many were earning less for similar orders.

An organizer collected these screenshots, and the team built a simple texting bot that read them with OCR and logged pay details into a shared spreadsheet. About 200 workers signed up within two weeks. The analysis showed that a large group had taken pay cuts, which supported a report and further protests. The talk then covers related projects: a phone app that gathers sensor data stored privately on the worker's device, spreadsheet templates for reconstructing unpaid hours, and digital workshops with unions in the Global South. A worker-led campaign also used the app to document unsafe late-night commutes.

Dan describes the obstacles, including employer wearables that collect data workers cannot access, often treated as trade secrets. He points to California's AB 701, which lets third parties request warehouse quota data on workers' behalf, and to data trusts and cooperatives as collective governance models. He closes with a 1913 long-exposure photograph of workers' motion, urging designers to keep people visible in data systems. The takeaways: work alongside worker groups, build tools that workers can adapt, and keep control of data with the workers who produce it.