When Humanlike AI Feels Wrong: Perceived Eeriness and Trust in AI-Mediated Brand Communication — A Systematic Literature Review
Inkubis : Jurnal Ekonomi dan Bisnis · 21.11.2026 · 8 (4) · 3603-3618
DOI: 10.59261/inkubis.v8i4.583 ↗Kurz gefasst
Systematische Übersicht in der Zeitschrift Inkubis : Jurnal Ekonomi dan Bisnis, online erschienen am 21.11.2026. Verfasst von Ignasius Bayu Purnomo und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-ND 4.0).
Abstract (Originalsprache)
Background
This review examines when humanlike artificial intelligence (AI) becomes unsettling and how uncanny appraisal relates to consumer trust in AI-mediated brand communication. It integrates the trust-building logic of anthropomorphism with evidence of unfavorable trust-related evaluations under uncanny conditions.
Methods
Following PRISMA 2020 and SPAR-4-SLR principles, a Scopus search identified 7,665 records. After duplicate, document-type, and title/abstract screening, 18 journal articles were retained, verified in full text, and coded by AI form, appraisal mechanism, trust object, boundary condition, outcome, and evidence directness. Additional theoretical sources were used for conceptual anchoring but were not added to the systematic evidence corpus.
Results
Five studies provide direct uncanny-appraisal evidence, six proximal-appraisal evidence, and seven contextual anthropomorphism–trust evidence. Direct studies associate unfavorable uncanny appraisal with lower trust-related evaluation, using distinct measures of affect, trust, and source trustworthiness. Human-likeness is not inherently harmful: social cues can support trust when aligned with competence, empathy, usefulness, transparency, and task demands.
Conclusion
The review develops the Eeriness-Trust Calibration Framework (ETCF), separating design cues from consumer appraisals, positive and negative trust routes, and multiple trust objects. It reframes AI humanization as a cue-capability-context calibration problem and identifies cross-object trust transfer to organizations and brands as a key unresolved marketing mechanism. Brands should optimize coherence rather than maximize realism, pretest eeriness and authenticity together, and monitor trust separately at agent, content, provider, and brand levels.
Zitieren
APA
Purnomo, I. B., Abadi, F., Sofia, R. D., & Edastama, P. (2026). When Humanlike AI Feels Wrong: Perceived Eeriness and Trust in AI-Mediated Brand Communication — A Systematic Literature Review. Inkubis : Jurnal Ekonomi dan Bisnis, 8(4), 3603-3618. https://doi.org/10.59261/inkubis.v8i4.583
Vancouver
Purnomo IB, Abadi F, Sofia RD, Edastama P. When Humanlike AI Feels Wrong: Perceived Eeriness and Trust in AI-Mediated Brand Communication — A Systematic Literature Review. Inkubis : Jurnal Ekonomi dan Bisnis. 2026;8(4):3603-3618. doi:10.59261/inkubis.v8i4.583
BibTeX
@article{Purnomo2026,
author = {Purnomo, Ignasius Bayu and Abadi, Ferryal and Sofia, Regina Deka and Edastama, Primasatria},
title = {When Humanlike AI Feels Wrong: Perceived Eeriness and Trust in AI-Mediated Brand Communication — A Systematic Literature Review},
journal = {Inkubis : Jurnal Ekonomi dan Bisnis},
year = {2026},
volume = {8},
number = {4},
pages = {3603--3618},
doi = {10.59261/inkubis.v8i4.583}
}
EUVIMED