TLDRocket
Sign in

AI Hallucination & Reliability

2 summarised stories about AI Hallucination & Reliability, each linking back to the original source. Browse all topics →

Friday, 17 July 2026

How Artificial Intelligence LLM Engines Shape the Global Conflict Information Environment

arXiv cs.AI 6 hours ago

Researchers tested five leading AI answer engines on questions about 28 conflicts and found they hallucinate more when conflicts have sparse documentation available. The engines made errors on 5,460 answers, with thinner records correlating to more invented details and misattribution. This vulnerability enables actors to manipulate AI responses through search optimization techniques, a practice already occurring on 1,048 websites analyzed, requiring policymakers to invest in local monitoring and translation-based research.

WavePhaseNet: A DFT-Based Method for Constructing Semantic Conceptual Hierarchy Structures (SCHS)

arXiv cs.CL 6 hours ago

Researchers propose WavePhaseNet, a method using Discrete Fourier Transform to construct semantic hierarchy structures in language models by decomposing embeddings into frequency bands and reducing dimensionality from 24,576 to approximately 3,000 dimensions. The approach achieves this 92% reduction in embedding dimensions while preserving semantic meaning by leveraging spectral analysis and Zipf's law. The method applies cohomological regularization to control consistency and suppress hallucinations through graph-based loss functions.