Experimental Research · Snap · 2025 · 10 weeks

The AI label doesn't hurt the brand. A bad ad does.

3
studies
167
total participants
66%
want AI transparency

The ask

Does knowing an ad was made by AI change how consumers feel about the brand — and does empathy explain why?

My role

Researcher · Analyst · Presenter

Timeline

2025 · 10 weeks

Skills used in this study — tap to explore

Consumers don't reject AI. They reject low-quality output. A credible ad survives disclosure. A non-credible ad fails regardless of who made it.

Key finding · AI in Advertising · 2025

Key findings

Finding 01Supported

The AI label alone had no significant effect on brand perception

Across two pre-registered experiments with different brands, simply labeling an ad as AI-created did not significantly change how consumers felt about the brand. The myth that disclosure automatically hurts brands is not supported by the data.

No significant main effect of AI label across Studies 1 and 2

Finding 02Supported

Ad credibility is the real variable

Higher credibility predicted higher brand perception regardless of whether the ad was human or AI-generated. Credibility is the safety net. Consumers don't reject AI; they reject the low-quality output associated with it.

Significant main effect of ad credibility on brand perception · Studies 1 and 2 combined

Finding 03Supported

Credibility moderates the AI label effect

A significant interaction showed that when credibility was low, the gap between human and AI ads was largest. As credibility increased, the difference diminished. A credible ad survives AI disclosure. A non-credible ad fails regardless.

Significant interaction: Ad condition × Ad credibility, p < .05

Finding 04Qualitative

AI acceptability is context-dependent

Study 3 qualitative coding revealed two tiers of AI acceptability. Transport, work, advertising, and medicine were rated higher acceptability. Politics, school, and creativity rated lower. 66% of respondents want transparency when brands use AI in advertising.

N=82 · 66% want transparency with AI usage in advertising

AI acceptability by context — mean rank (1 = most acceptable, 7 = least)

Lower rank = more acceptable · N=41

Tier 1 — higher acceptability
Transportation
3.00
Workplace
3.02
Advertising
3.22
Medicine
3.44
Tier 2
Schools
4.52
Creative Arts
5.37
Politics
5.37

Friedman χ²(6) = 61.06, p < .001 · N=41 · No significant differences within tiers

Study 3 · Qualitative data · N = 82

Open-ended coding on AI sentiment and the transparency preference breakdown.

AI sentiment themes

Top 3 themes from qualitative coding of open-ended responses

Neutral toolAI as practical, efficiency-supporting
15%
OverrelianceConcern about loss of human thinking
12%
Harmful to peopleStrips agency, threatens livelihoods
11%

Perception is mixed — neither purely negative nor positive

Remaining ~62% of coded sentiments span approval, curiosity, and ambivalence

AI transparency preference

Would you want to know if your favorite brand used AI in their ads? (N=82)

66% — want to know, to stop supporting the brand or because they value transparency
34% — neutral or no stated preference

Study 3 · Open survey question · N=82

Research process

Tap a card to flip it. Use arrows to move through the deck.

01 / 06

01
tap to flip ↻

AI disclosure · narrative advertising · empathy

Literature review

Existing research on AI in advertising measured attitude toward AI generally, not toward specific AI-labeled content. Nobody had tested whether empathy mediates the relationship between AI awareness and brand perception.

Literature reviewtap to flip ↻

What I did

Existing research on AI in advertising measured attitude toward AI generally, not toward specific AI-labeled content. Nobody had tested whether empathy mediates the relationship between AI awareness and brand perception.

Why

Without a theoretical mechanism, you can measure an effect but can't explain it. Identifying empathy as a potential mediator gave the study a direction.

What it unlocked

A causal model: AI awareness → empathy → brand perception, with ad credibility as a moderator.

Get your results

Want to know what your condition revealed about how you respond to design mechanics?

No spam. One email with your personalized findings.

Next project

What happens when the portfolio itself is the research instrument?