{
  "id": 121,
  "url": "https://arxiv.org/abs/2607.07605v1",
  "title": "User identity conditions moral wrongness ratings in non-reasoning large language models",
  "summary": "This study adopts a behavioural bottom-up approach to AI value alignment to investigate whether an implicitly conveyed user identity shifts the moral evaluations of large language models (LLMs). Through a structured, multi-turn conversational protocol across 12,000 interactions, we evaluate AI value alignment in two non-reasoning models, gpt-4.1-mini-2025-04-14 and gemini-2.5-flash-lite. Rather than instructing the models to adopt a persona or prompting them with explicit moral stances, the user",
  "authors": "Willem Fourie, Isabel Ray, Gray Manicom",
  "category": "research",
  "topics": "safety-alignment,finance-investment",
  "orgs": null,
  "regions": null,
  "published_at": "2026-07-08T16:22:27.000Z",
  "fetched_at": "2026-07-14T14:14:19.967Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/121",
  "original_url": "https://arxiv.org/abs/2607.07605v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}