Signed-Graph Recommendation as Structural Consistency Maximization
2026-07-08T08:52:19Z•dcab4d9698b74e691186c76b3d4525c094f2d9bf727ef55446712878dea0bed7
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.