Navigating the AI Revolution in the Era of Community-Accountable Archaeological Praxis

Abstract

Abstract This article explores how generative artificial intelligence (AI), particularly Large Language Models (LLMs) such as OpenAI’s ChatGPT, can be mobilized to support collaborative and community-accountable archaeological practice. We situate this exploration within the growing conversation about the use and utility of generative AI in archaeological practice and a discussion about the capabilities of LLMs. Through three case studies—Picuris Pueblo’s legacy collections documentation; the transcription of early twentieth-century archival materials in San Ildefonso Tultepec, Querétaro; and the digitization of the Mountain Messenger —we evaluate the potentials and pitfalls of LLMs in supporting document transcription, organization, and analysis within community-led research contexts. Our findings suggest that, when used transparently and under community direction, LLMs can potentially increase efficiency and accessibility in data processing, expand opportunities for capacity building, and promote citizen-science approaches to heritage stewardship. However, we also identify critical challenges, including data sovereignty, environmental concerns, inequitable access to digital infrastructure, and the persistent need for human oversight to ensure accuracy and accountability.