CLHCApr 17

LLMs Corrupt Your Documents When You Delegate

arXiv:2604.1559727.72 citationsh-index: 4
Predicted impact top 40% in CL · last 90 daysOriginality Incremental advance
AI Analysis

For users relying on LLMs for delegated document editing, this paper reveals a critical reliability issue that undermines trust in current AI systems.

The paper introduces DELEGATE-52, a benchmark simulating long delegated document editing workflows across 52 domains, and finds that current LLMs corrupt an average of 25% of document content by the end of such workflows, with errors compounding over time.

Large Language Models (LLMs) are poised to disrupt knowledge work, with the emergence of delegated work as a new interaction paradigm (e.g., vibe coding). Delegation requires trust - the expectation that the LLM will faithfully execute the task without introducing errors into documents. We introduce DELEGATE-52 to study the readiness of AI systems in delegated workflows. DELEGATE-52 simulates long delegated workflows that require in-depth document editing across 52 professional domains, such as coding, crystallography, and music notation. Our large-scale experiment with 19 LLMs reveals that current models degrade documents during delegation: even frontier models (Gemini 3.1 Pro, Claude 4.6 Opus, GPT 5.4) corrupt an average of 25% of document content by the end of long workflows, with other models failing more severely. Additional experiments reveal that agentic tool use does not improve performance on DELEGATE-52, and that degradation severity is exacerbated by document size, length of interaction, or presence of distractor files. Our analysis shows that current LLMs are unreliable delegates: they introduce sparse but severe errors that silently corrupt documents, compounding over long interaction.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes