Scholarlink

The process of obtaining user consent for personal data management has become a significant burden for software developers. The difficulty increases as consent-related requirements must be continuously updated to accommodate new templates and regulatory specifications. Recent advances in Large Language Models (LLMs) offer promising potential to automate this process. In this paper, we examine the capabilities of popular LLMs (LLaMa, Gemma, DeepSeek, and Phi-3) to generate consent requirements in collaboration with developers providing natural language inputs. We conducted an empirical, multi-dimensional evaluation encompassing both consent form quality and developer productivity. The findings reveal a trade-off between model complexity, output quality, and productivity, exacerbated by the context of the requirement. The paper provides useful implications for practitioners selecting LLMs for consent engineering.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Evaluating Large Language Models for Consent Engineering in Data-Driven Applications

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Anastasia Terzi,
  • Christina Zoi,
  • Stamatia Bibi,
  • Pantelis Angelidis,
  • Markos Tsipouras

摘要

The process of obtaining user consent for personal data management has become a significant burden for software developers. The difficulty increases as consent-related requirements must be continuously updated to accommodate new templates and regulatory specifications. Recent advances in Large Language Models (LLMs) offer promising potential to automate this process. In this paper, we examine the capabilities of popular LLMs (LLaMa, Gemma, DeepSeek, and Phi-3) to generate consent requirements in collaboration with developers providing natural language inputs. We conducted an empirical, multi-dimensional evaluation encompassing both consent form quality and developer productivity. The findings reveal a trade-off between model complexity, output quality, and productivity, exacerbated by the context of the requirement. The paper provides useful implications for practitioners selecting LLMs for consent engineering.