Studying Generative AI at Work: Participatory Case Studies for Building Labor Resistance

• In Event: Labor and Technology: Platforms, AI, and Inequalities

Abstract

Workers have been fighting technological control in their workplace for as long as bosses have thought to use it. In the current era, workers are pushing back against the new rash of developments in computing technology, including generative artificial intelligence (AI) tools, such as ChatGPT and Midjourney; workplace surveillance technologies which collect data on workers without their consent; and algorithmic management tools which are often trained on workers’ own data and used to highly constrain their autonomy. Those who have opposed the rapid adoption and integration of AI systems, and pushed for greater regulation and oversight, are often characterized as irrational fear-mongerers, stifling innovation, or Luddites (at least in a pejorative register). Despite these challenges, workers have successfully integrated AI issues into their broader contract campaigns, winning important AI protections in their workplaces while also bringing greater public attention to the harms caused by the development and use of these technologies.

This paper presents our learnings as part of a worker-centric project, the Luddite Lab. It draws on insights from three labor unions based in the United States, representing workers from three different sectorshealthcare, journalism, and information technology- to create applied research for worker groups fighting automation at work. These case studies were developed through participatory action research with the key members of the labor unions and outline cross-sector insights on successful strategies for worker control of automation in the workplace. After initial review of the in-depth session notes across twelve sessions with three unions, we identified three important insights relevant for other unions and worker collectives resisting AI in their workplaces: (1) fixating solely on AI can be a “red herring”; (2) workers best understand how to incorporate new tools into their roles; and (3) political education is critical for inoculation against misinformation.

Authors

• Tina M Park, The Distributed AI Research Institute

• Alex Hanna, Distributed AI Research Institute

• Sophie Song, Distributed Artificial Intelligence Research Institute