Signed-Graph Recommendation as Structural Consistency Maximization

2026-07-08T08:52:19Zdcab4d9698b74e691186c76b3d4525c094f2d9bf727ef55446712878dea0bed7
LLM agentsalgorithmic biascitation fraudcontrastive learningdisplacement estimationgraph machine learninginformation manipulationlarge language modelsmisinformationmobile-phone datamodel hallucinationmulti-agent systemspreferential attachmentprivacyre-identification riskrecommendation systemsscholarly-metric manipulationsemantic intent detectionsigned graphssocial-simulation

What happened

This feed aggregates recent arXiv CS/social-impacts research with several security-relevant themes. Key items: (1) studies of autonomous LLM agent networks show emergent preferential-attachment and type-dependent "glass-ceiling" centrality effects, implying delegation and collaboration dynamics can produce persistent structural disparities that adversaries or low-quality models could exploit to gain undue influence in multi-agent workflows; (2) audits of LLM restaurant recommendations reveal systematic fabrication (hallucination) and selective invisibility of real venues, plus user-targeted (s

Why it matters

A reviewed impact interpretation has not been published for this record.

Evidence and limitations

Source ID
arxiv_cs_si
Record identifier
dcab4d9698b74e691186c76b3d4525c094f2d9bf727ef55446712878dea0bed7
Enrichment time
2026-07-08T08:52:19Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.

Record · Signed-Graph Recommendation as Structural Consistency Maximization · Baitaphish