Research & Thinking
Research & Thinking · 001

Human or Machine?

What consumer perceptions of AI-generated advertising mean for marketers.

This research explored how consumers evaluated advertising they perceived as AI-generated versus human-created across creativity, emotional engagement, authenticity, trust and overall attitude toward the advertisement.

BSc Marketing Dissertation · May 2026
Sample
n = 86
Stimuli
6 ads
Brand
Lumera (fictitious)
Method
Online survey
01The Commercial Question

AI can produce more advertising, faster. But can it create the same emotional meaning and trust as human creative work?

Generative AI has collapsed the cost of producing advertising. Brands can now generate variations, visuals and copy at a scale that was previously impossible. The commercial question is no longer whether AI can make an ad — it clearly can.

The question is what happens to the perceptions that make advertising work. Creativity, emotion, authenticity and trust are not incidental — they are the mechanism through which advertising builds brand equity. If consumers evaluate AI-associated advertising differently, that has direct implications for how brands deploy it.

02Research Design

A controlled consumer study, honestly reported.

Fictitious skincare brand
Lumera
Advertising stimuli
Six ads
Instrument
Online consumer survey
Final analytical sample
86 participants
Measured constructs
Creativity · Emotion · Authenticity · Trust · Ad attitude
Methodological note

The original experimental design was adapted after the condition variable was unavailable. The final analysis therefore compared consumers based on the source they believed created the advertisement. The findings relate to perceived authorship rather than proving direct causal differences between actual AI- and human-created ads.

03Key Findings

Ads perceived as human-created were rated meaningfully higher — across every construct.

Reported as Cohen's d — the standardised difference between groups. In this sample, all five effects were large by conventional benchmarks. Findings apply to this specific sample and skincare advertising context.

Perceived human vs. AI
d=0.00
Emotional engagement
Large effect
Perceived human vs. AI
d=0.00
Creativity
Large effect
Perceived human vs. AI
d=0.00
Advertising attitude
Large effect
Perceived human vs. AI
d=0.00
Authenticity
Large effect
Perceived human vs. AI
d=0.00
Trust
Large effect

Note: Effect sizes describe the magnitude of the group difference observed in this study. They are not claims about all consumers or all advertising categories.

04The Central Insight

The problem was not simply whether AI could make a polished advertisement.

It was whether consumers believed genuine human meaning existed behind it.

Consumers appeared to associate human authorship with greater intentionality, emotional investment and authenticity. These perceptions shaped trust and, ultimately, attitudes toward the advertisement.

05Authenticity & Trust Pathway

How perception becomes attitude.

Stage 1
Perceived source
Stage 2
Perceived authenticity
Stage 3
Trust
Stage 4
Attitude toward the advertisement

A serial mediation analysis showed authenticity and trust operating sequentially between perceived source and advertising attitude. The model explained a large share of the variance in attitude toward the advertisement — but this is a statistical association within the sample, not proof of universal causality.

73.9%
Variance explained · R²
06Strategic Implications for Marketers

Four practical decisions the evidence points toward.

Recommendation 01

Use AI for scale

Variation, adaptation, ideation and production efficiency. AI shortens the distance between idea and executed asset.

Recommendation 02

Keep humans responsible for meaning

Strategy, cultural judgment, emotional nuance and brand interpretation remain hard to automate — and are what consumers respond to.

Recommendation 03

Protect authenticity

Do not allow automation to make brand communication feel impersonal, opportunistic or generic. Authenticity is where trust is built.

Recommendation 04

Match AI use to context

AI may be more suitable for functional communications than emotionally sensitive, identity-led or high-trust categories.

The strongest model is not human versus machine.

It is human judgment amplified by AI capability.

07Research Limitations

What this study can — and cannot — claim.

Rigour requires stating the boundaries of a study as clearly as its findings. The results should be read as directional, not universal.

  • L/01

    The final analysis was based on perceived rather than actual source condition — findings speak to how consumers respond to what they believe about authorship.

  • L/02

    The study used a skincare advertising context; emotionally sensitive or identity-led categories may behave differently to functional ones.

  • L/03

    The final analytical sample of 86 participants is limited and skewed toward a younger, digitally native demographic.

  • L/04

    Effects are statistical associations within this sample. They are best interpreted as directional evidence, not universal laws of consumer behaviour.

08Closing Reflection

AI may increasingly outperform humans in speed, scale and executional volume.

But where advertising depends on authenticity, trust and emotional meaning, human judgment remains strategically valuable.