The rise of Large Language Models (LLMs)—AI systems trained on vast amounts of text data that can understand and generate human language—is unsettling conventions in anthropology and archaeology education, adding to this Special Issue’s call for building post-conventional practice. Attending to the effects of emerging technologies on education, we note opportunities for LLMs to support learning without dulling scholarly effort or reducing AI engagements to morality debates. We argue that for students to effectively use LLMs it is important that they can first critique AI in the context of coursework. Undocumented labour, a key concern of AI use at university, mirrors the discipline’s own reckoning with ‘fieldwork wives’ and unacknowledged field assistants, offering opportunities to explore knowledge co-production historically. To build critical AI and disciplinary practice, we put forward the PRIDE model—Prompt, Reflection, Implementation, Discernment, and Evaluation—and discuss an assessment designed to supplement students’ understanding of disciplinary debates and foundational concepts. The task required students to select an LLM, develop promoting strategies, interact to co-produce knowledge, and then critically appraise this interaction. Surprise, curiosity and intrigue were noted as guiding affects through which students reflected on their experience. We discuss the potential for LLMs to introduce students to socio-cultural complexity through affective encounters. Finally, we provide an example assessment and rubric to encourage broader uptake of AI-mediated assessments to build critical skills in AI use.
Heffernan et al. (Mon,) studied this question.