Anna Neumann
Anna Neumann is a PhD candidate in Computer Science at the Research Center for Trustworthy Data Science and Security. Her research focuses on transparency and governance requirements of technical systems to give stakeholders meaningful agency and recourse over systems, which has been recognized with a CHI Best Paper Award.
She is part of the Compliant and Accountable Systems Group, which also focuses on perceptions of technological systems towards better governance of them. Anna has organized workshops, retreats, and mini-conferences for the RC Trust Graduate School which encompasses around 35 PhD students.
Affiliation: Research Center for Trustworthy Data Science and Security, University Duisburg-Essen.
Who Controls the Conversation? User Perspectives on Generative AI (LLM) System Prompts
Anna Neumann, Yulu Pi, Jatinder Singh
Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems
System prompts - instructions that shape the behaviour of generative AI systems - strongly influence system outputs and users' experiences. They define the model's guidelines, `personality', and guardrails, taking precedence over user inputs. Despite their influence, transparency is limited: system prompts are generally not made public and most platforms instruct models to conceal them, leaving users disconnected from and unaware of a key mechanism guiding and governing their AI interactions. This paper argues that system prompts warrant explicit, user-centred design attention and, focusing on large language models (LLMs), asks: what do system prompts contain, how do end-users perceive them, and what do these perceptions offer for design and governance practice? Our results reveal user perspectives on: the benefits and risks of system prompts; the values they prefer to be associated with prompt-design; their levels of comfort with different types of prompts; and degrees of transparency and user control regarding prompt content. From these findings emerge considerations for how designers can better align system prompt mechanisms with user expectations and preferences over these mechanisms that directly shape how generative AI systems behave.
