Metraipolis › archive › dc_6b929057eaaf
Evaluating the Hidden Costs of Personalization in Large Language Models — arXiv:2608.28833
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.28833
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T07:43:52.929Z
- SHA-256 of the archived text
2bc1c4ea38f076bff7eff9e77fbecb71d38d13d9d73109b44102ea4da64fb7fd- Stable URL of this page
https://metraipolis.com/documento/dc_6b929057eaaf
How to cite
Evaluating the Hidden Costs of Personalization in Large Language Models — arXiv:2608.28833 — arXiv (en), https://arxiv.org/abs/2608.28833 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6b929057eaaf (Metraipolis archive dc_6b929057eaaf, sha256 2bc1c4ea38f076bff7eff9e77fbecb71d38d13d9d73109b44102ea4da64fb7fd)
BibTeX
@misc{dc6b929057eaaf,
title = {Evaluating the Hidden Costs of Personalization in Large Language Models — arXiv:2608.28833},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.28833},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6b929057eaaf; sha256 2bc1c4ea38f076bff7eff9e77fbecb71d38d13d9d73109b44102ea4da64fb7fd},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.28833. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
While Large language models (LLMs) incorporate user personalization signals to improve usability and helpfulness, they increasingly shift from providing balanced, informative responses toward optimizing for user satisfaction when conditioned on personal context such as conversation history, inferred preferences, and user profiles. Specifically, we identify three emerging risks: (1) irrelevant personalization, where models reference personal information in unnecessary contexts; (2) preference narrowing, where models reinforce informational echo chambers; and (3) sycophantic bias, where models agree excessively with user opinions. As a result, models may reference personal information in contexts where it is unnecessary, inadvertently collapse response diversity, or agree excessively with user opinions. Despite the growing use of personalization in AI assistants, there has been limited systematic evaluation of its potential side effects. To bridge this gap, we propose PRISK, a dynamic evaluation framework with automated data generation and tailored metrics that uncovers systematic limitations in current LLM personalization and how personalized information shapes its responses. Our empirical analysis across 13 LLMs demonstrates the presence of user profiles and retrieved memories consistently exacerbates biases, resulting in an average drop of 45.9% in irrelevant personalization, 41.7% in preference narrowing and 61.7% in sycophantic bias. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-01 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.28833 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-17T07:28:26.485Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).