First-pass research briefing, not a final academic review. Always read the original paper before citing.
Paper A
Generative AI in Advertising: Consumer Responses to AI-Disclosed Ad Copy
Johnson, M., Chen, L., Patel, R. — 2025 — Journal of Marketing Research
peer reviewed journal article · abstract only · deep dive
https://doi.org/10.1177/00222437251234567
Key findings
- Consumers rate AI-disclosed ads 12% lower on trust but 8% higher on novelty
- Disclosure effects are moderated by brand familiarity
- High-involvement categories show stronger negative disclosure effects
Marketing implications
- Test AI disclosure wording before rolling out to full campaigns
- Disclosure matters more for financial and health products
Paper B
Algorithm Aversion in Recommendation Systems: When Consumers Reject Accurate AI Suggestions
Williams, S., Okonkwo, A. — 2025 — Journal of Consumer Psychology
peer reviewed journal article · open access available · deep dive
https://doi.org/10.1016/j.jcps.2025.03.002
Key findings
- Consumers reject AI recommendations even when accuracy is superior
- A single failure doubles long-term algorithm aversion
- A small override option reduces aversion by 34%
Marketing implications
- Always include a human override option in recommendation UX
- One bad recommendation can permanently reduce engagement with AI features
Paper C
Virtual Influencers and Brand Authenticity: A Meta-Analysis
Martinez, E., Kim, J., Thompson, B., Lee, C. — 2024 — International Journal of Advertising
meta analysis · abstract only · skim later
https://doi.org/10.1080/02650487.2024.2198765
Key findings
- Virtual influencers generate comparable engagement to humans for tech products
- Authenticity gap narrows significantly for Gen Z
- Brand fit moderates effectiveness more than influencer type
Marketing implications
- Virtual influencers are viable for tech and gaming with Gen Z
- Brand fit matters more than whether the influencer is human or virtual
Paper D
AI Personalization and Privacy: The Creepiness Threshold in Targeted Advertising
Brown, K. — 2025 — Marketing Science
peer reviewed journal article · paywalled · deep dive
https://doi.org/10.1287/mksc.2025.0123
Key findings
- Personalization crosses the creepiness threshold at ~70% accuracy on unstated preferences
- Temporal distance in data collection reduces perceived intrusiveness
- Opt-in framing reduces creepiness by 41% vs opt-out
Marketing implications
- Avoid hyper-precise signals that reveal inferred preferences
- Default to opt-in consent flows for high-sensitivity targeting
Paper E
LLM-Generated Ad Copy vs Human Copywriters: Persuasion, Recall, and Purchase Intent
Garcia, F., Nakamura, Y., Singh, P. — 2025 — Journal of Advertising
peer reviewed journal article · full text available · deep dive
https://doi.org/10.1080/00913367.2025.2156789
Key findings
- LLM copy matches human copy on recall but scores 9% lower on emotional resonance
- Hybrid human-LLM copy outperforms both on purchase intent
- Effect size varies by product category and audience segment
Marketing implications
- Use LLMs for volume copy, humans for emotional and brand-voice refinement
- Test hybrid workflows before fully replacing copywriters with AI
AI & Marketing Research Radar — Big Plans Media — 2026-05-06
