Renegotiating Library Values: Black LIS Professionals’ Critical Refusal of Generative AI

With the rapid expansion of large language model (LLM) capabilities, librarians, especially in academic libraries, are grappling with how to use AI technology while maintaining library values of providing accurate and reliable information for all. Generative AI poses several challenges to privacy, voice, intellectual freedom, and the environment. For LIS workers, generative AI can pose threats to expectations of labor, valuing technological skill over the emotional and reproductive labor of librarianship. Additionally, race and gender bias reinforce and reinvent stigma and discrimination through the use of LLMs. Critical refusal offers a lens through which LIS professionals can reimagine power and autonomy in library decision-making. Through semi-structured interviews with 7 Black LIS professionals who refuse to use generative AI in their work, this study examines why they refuse it in the workplace, how refusal informs conversations on AI ethics, and the methods and implications of refusal in libraries. The study found that refusal, while manifested in a variety of ways for a variety of reasons, can be a useful mechanism for the LIS profession to identify the needs and concerns of those most marginalized, and work towards adopting more equitable, less harmful technologies in libraries and archives.