{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,10,16]],"date-time":"2025-10-16T00:47:03Z","timestamp":1760575623884,"version":"build-2065373602"},"reference-count":0,"publisher":"Association for the Advancement of Artificial Intelligence (AAAI)","issue":"1","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["AIES"],"abstract":"<jats:p>We study how partial information about scoring rules\naffects fairness in strategic learning settings. In\nstrategic learning, a learner deploys a scoring rule, and\nagents respond strategically by modifying their\nfeatures---at some cost--\u2014to improve their outcomes.\nHowever, in our work, agents do not observe the scoring\nrule directly; instead, they receive a noisy signal\nof said rule. We consider two different agent models: (i)\nnaive agents, who take the noisy signal at face\nvalue, and (ii) Bayesian agents, who update a prior\nbelief based on the signal.\n\nOur goal is to understand how disparities in outcomes arise\nbetween groups that differ in their costs of feature\nmodification, and how these disparities vary with the\nlevel of transparency of the learner's rule. For\nnaive agents, we show that utility disparities can grow\nunboundedly with noise, and that the group with lower costs\ncan, perhaps counter-intuitively, be disproportionately\nharmed under limited transparency. In contrast, for\nBayesian agents, disparities remain bounded. We provide a\nfull characterization of disparities across groups as a\nfunction of the level of transparency and show that they\ncan vary non-monotonically with noise; in\nparticular, disparities are often minimized at\nintermediate levels of transparency. Finally, we\nextend our analysis to settings where groups differ not\nonly in cost, but also in prior beliefs, and study how this\nasymmetry influences fairness.<\/jats:p>","DOI":"10.1609\/aies.v8i1.36544","type":"journal-article","created":{"date-parts":[[2025,10,15]],"date-time":"2025-10-15T13:15:53Z","timestamp":1760534153000},"page":"227-237","source":"Crossref","is-referenced-by-count":0,"title":["The Disparate Effects of Partial Information in Bayesian Strategic Learning"],"prefix":"10.1609","volume":"8","author":[{"given":"Srikanth","family":"Avasarala","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Serena","family":"Wang","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Juba","family":"Ziani","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"9382","published-online":{"date-parts":[[2025,10,15]]},"container-title":["Proceedings of the AAAI\/ACM Conference on AI, Ethics, and Society"],"original-title":[],"link":[{"URL":"https:\/\/ojs.aaai.org\/index.php\/AIES\/article\/download\/36544\/38682","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/ojs.aaai.org\/index.php\/AIES\/article\/download\/36544\/38682","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,15]],"date-time":"2025-10-15T13:15:54Z","timestamp":1760534154000},"score":1,"resource":{"primary":{"URL":"https:\/\/ojs.aaai.org\/index.php\/AIES\/article\/view\/36544"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,15]]},"references-count":0,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2025,10,15]]}},"URL":"https:\/\/doi.org\/10.1609\/aies.v8i1.36544","relation":{},"ISSN":["3065-8365"],"issn-type":[{"value":"3065-8365","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,10,15]]}}}