Enterprise AI assistants suffer from two failure modes not captured by existing public-web QA benchmarks: authorized information leakage, where factually correct answers reach unintended readers, and aggregation failure, where evidence scattered across artifacts yields fluent but incomplete responses. Both stem from enterprise authority being per-item and partly body-embedded, a structure that existing benchmarks address along at most one axis. We argue that authority-aware evaluation must be conducted at the per-item level, and instantiate this principle through Cross-Artifact and Authority-aware Reasoning over Organizational Bundles (CAROB): a synthetic enterprise dataset in which every retrieval-and-citation item carries a time-indexed audience/approval/lifecycle annotation, paired with three tasks that decouple cross-artifact coherence, authority-aware retrieval, and role-conditioned answering. Evaluating five frontier and open-weight LLMs, we find that no single model dominates both axes of the benchmark: each task induces a distinct model ordering, and the model strongest on authority handling ranks last on answer coverage. Two structural patterns hold across the lineup: retrieval-side leakage amplifies substantially at the answer stage, and structural authority filtering eliminates the residual leak band that prompt-level mitigation leaves behind, with both patterns replicating on EnronQA and ACCESS DENIED INC. We release CAROB as an evaluation testbed for authority-aware enterprise system design, together with its per-item annotation, three-task structure, and label-first generation pipeline as reusable infrastructure.

Continual pre-training (CPT) has been widely adopted as a method for domain expansion in large language models. However, CPT has consistently been accompanied by challenges, such as the difficulty of acquiring large-scale domain-specific datasets and high computational costs. In this study, we propose a novel method called Test-Enhanced Learning for Language Model Enrichment (TELLME) to alleviate these issues. TELLME leverages the Test-Enhanced Learning (TEL) principle, whereby the model’s learning efficiency is improved using quizzes during training. It integrates this principle with CPT, thereby promoting efficient domain-specific knowledge acquisition and long-term memory retention. Experimental results demonstrate that TELLME outperforms existing methods by up to 23.6% in the financial domain and achieves a 9.8% improvement in long-term memory retention.

Large language models (LLMs) can generate human-like online comments at scale, raising concerns that coordinated AI campaigns could manipulate public opinion. While prior work has shown that single AI messages can persuade individual readers, it remains untested whether coordinated multi-agent LLM campaigns, in which multiple agents jointly plan, divide roles, and manage visibility, can shift reader opinions in news comment sections, and whether per-comment AI labels can mitigate such influence. We address these questions through a controlled experiment with 1,000 U.S. adults reading news articles on a Reddit-like platform. Each participant was randomly assigned to one of six conditions, crossing three comment sources (human, independently generated AI, or coordinated multi-agent LLM) with two label settings (per-comment "Likely AI-generated" badges from a deployed detector, shown or hidden). Coordinated LLM comments did not increase average persuasion over independently generated AI comments, but disproportionately converted readers who initially disagreed and compressed opinion diversity by 12%, manufacturing apparent consensus where genuine disagreement existed. The labels reduced the perceived believability of comments but did not reliably reduce actual opinion change. These findings reveal what we call a Knowing-Resisting dissociation between source recognition and behavioral resistance, and indicate that per-comment disclosure alone is insufficient to counter coordinated LLM opinion campaigns.

Large language models (LLMs) generate human-like text, raising concerns about their misuse in creating deceptive content. Detecting LLM-generated comments (LGC) in online news is essential for preserving online discourse integrity and preventing opinion manipulation. However, effective detection faces two key challenges; the brevity and informality of news comments limit traditional methods, and the absence of a publicly available LGC dataset hinders model training, especially for languages other than English. To address these challenges, we propose a twofold approach. First, we develop an LGC generation framework to construct a high-quality dataset with diverse and complex examples. Second, we introduce XDAC (XAI-Driven Detection and Attribution of LLM-Generated Comments), a framework utilizing explainable AI, designed for the detection and attribution of short-form LGC in Korean news articles. XDAC leverages XAI to uncover distinguishing linguistic patterns at both token and character levels. We present the first large-scale benchmark dataset, comprising 1.3M human-written comments from Korean news platforms and 1M LLM-generated comments from 14 distinct models. XDAC outperforms existing methods, achieving a 98.5% F1 score in LGC detection with a relative improvement of 68.1%, and an 84.3% F1 score in attribution. To validate real-world applicability, we analyze 5.24M news comments from Naver, South Korea’s leading online news platform, identifying 27,029 potential LLM-generated comments.