Summary
Getting computer programs like large language models to give useful answers depends on how people ask questions. This paper says that the way prompts are written not only affects the answers but also what values and assumptions influence them. The authors explain two ways people use prompts: treating the model as a knowledge expert or feeding it facts to organize. They suggest the second is better for responsible use and give five rules for ethical prompting, including admitting uncertainty and knowing the system’s limits. They also say responsibility for AI output is shared among users, developers, and society, and better prompts alone can’t fix all ethical issues.
What this means in practice
- •For ai product teams: Design internal prompting guidelines based on principles that reduce misleading or biased outputs from large language models.
- •For content moderation teams: Develop prompt strategies that explicitly address uncertainty and avoid deceptive framing to ensure trustworthy content generation oversight.
A position paper. It proposes an approach and reports no results.
Abstract
Prompt engineering is commonly understood as a technical practice aimed at eliciting useful, accurate, and appropriately structured outputs from large language models (LLMs). Yet prompts do more than optimize model performance: they shape what information is elicited, which assumptions are introduced, and which perspectives are privileged. In this paper, we argue that prompting should therefore be understood as an epistemic and normative act. We distinguish between two conceptions of prompting: knowledge generation, in which the model is implicitly treated as an epistemic authority, and knowledge injection, in which users provide relevant knowledge, evidence, and context that the model can organize, articulate, and extend. We contend that the latter provides a more appropriate model for responsible LLM use, since prompting cannot substitute for subject-matter expertise or the critical evaluation of model outputs. At the same time, prompts inevitably inject values, intentions, and assumptions into the generation process. Building on these premises, we propose five principles for responsible prompting: epistemic humility, explicit engagement with uncertainty and epistemic underdetermination, source sensitivity, avoidance of deceptive framing, and recognition of model limitations and appropriate task allocation. We further argue that responsibility for AI-generated outputs is distributed across users, developers, institutions, and societies, and that some applications cannot be rendered ethical through better prompting alone.