Artificial Intelligence in Advertising: A Scientometric and Thematic Synthesis of Four Paradoxes

Adalı G., Yardibi F., Talay Ö.

JOURNAL OF INTERACTIVE ADVERTISING, cilt.26, ss.1-18, 2026 (ESCI, Scopus)

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 26
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1080/15252019.2026.2710678
  • Dergi Adı: JOURNAL OF INTERACTIVE ADVERTISING
  • Derginin Tarandığı İndeksler: Business Source Ultimate (EBSCO), Scopus, Emerging Sources Citation Index (ESCI), IBZ Online, Psycinfo
  • Sayfa Sayıları: ss.1-18
  • Akdeniz Üniversitesi Adresli: Evet

Özet

Artificial intelligence (AI) advertising research has been expanding rapidly but remains conceptually fragmented. Findings across the field pull in opposite directions: personalization improves relevance yet erodes trust, automation accelerates production yet compresses creative autonomy, and anthropomorphic AI agents can strengthen engagement yet provoke discomfort. Existing bibliometric and scientometric studies map what is being studied without explaining why findings contradict each other. Stage 1 of the current study, a scientometric analysis of 475 Web of Science publications (2005–2025), showed 38% average annual growth, a shift from technical optimization toward human-centric themes, and the dominance of the Human–AI Collaboration, AI-Generated Advertising, and Virtual Influencer clusters. Stage 2 of the current study applied thematic synthesis to 22 quality-assessed qualitative and mixed-method communication studies. The synthesis produced seven analytical themes organized around four structurally linked paradoxes: personalization versus surveillance, automation versus creative agency, anthropomorphism versus authenticity, and transparency versus cognitive overload. In each paradox, the mechanism producing the benefit was found to produce the cost as well; such tensions arise from the constitutive logic of AI-mediated advertising instead of poor implementation. The contradictions become structurally predictable, and practitioners gain decision frameworks for the tensions AI integration generates.