The European Union's regulations on sustainability and privacy require companies to create documentation artifacts. However, creating these artifacts is challenging due to industrial data being scattered across company and supplier systems in heterogeneous formats. Researchers have proposed using Large Language Models (LLMs) to generate compliance artifacts, but the impact of data extraction instructions and regulatory vagueness on LLM output is unclear. A study investigates how these factors affect the quality and consistency of LLM-produced compliance artifacts, finding that less strict guidelines require more context prompts for consistent results, while stricter guidelines may lead to hallucinations in the output. --- Why it matters: This research matters because it highlights the challenges of implementing EU regulations on sustainability and privacy using LLMs. Understanding how these models perform under different conditions can inform the development of more effective regulatory compliance tools. Source: https://arxiv.org/abs/2608.21317