Authors: Sylvana Sofkova Hashemi Source: Computers and Education: Artificial Intelligence, Vol 11, 100635 — Open Access (CC BY 4.0)
Key Findings
Empirical study with 28 postgraduate students engaged in collaborative multimodal prompting tasks. Found prompting strategies vary along a continuum from basic input-output to strategic, iterative, dialogic practices. Prompting emerges as a central epistemic practice — students critically interpret, refine, and negotiate AI outputs. Multimodal engagement exposes challenges in translating abstract meaning into machine-readable prompts. Ethical dimensions of AI literacy remain underdeveloped despite gains in evaluation and creation.