The Headline Finding
Research published in Harvard Business Review on September 3, 2026, found that AI-generated ads performed, on average, 14% weaker than human-made ads on short-term sales potential and 17% weaker on long-term brand equity measures. This was true even when consumers could not reliably distinguish AI-generated ads from human-made ones (Adam Peruta, HBR, September 3, 2026).
The study, conducted by Syracuse University's S.I. Newhouse School of Public Communications in collaboration with Ipsos, tested 20 AI-generated and human-made video ad pairs on approximately 3,000 consumers. The ads were matched on strategy, category, and format to isolate the creative difference.
Why This Matters Now
Generative AI has made creative production faster and cheaper. A team can now generate concepts, scripts, images, and even video spots in a fraction of the time it once took. The output often looks polished, on-brand, and real. But looking real enough to make consumers believe they are seeing human-designed ads is not the same as working well.
Adam Peruta, the study's lead researcher and associate professor at Syracuse University's Newhouse School, described the real risk: work that looks good enough to approve while underperforming on the metrics that actually drive business results. This is what makes the finding dangerous. It is not that AI ads are obviously bad. They are good enough to pass review while quietly leaving value on the table.
The Numbers That Should Make You Pause
| Measure | AI-Generated Ads | Human-Made Ads | Gap |
|---|---|---|---|
| Short-term sales potential | Baseline | 14% stronger | 14% |
| Long-term brand equity | Baseline | 17% stronger | 17% |
| Consumer differentiation | Could not reliably tell | Could not reliably tell | N/A |
| Strategy control | Matched to human | Matched to AI | Controlled |
The 14% gap on short-term sales potential means that if a human-made ad would generate 100 units of sales lift, the AI-generated equivalent would generate roughly 86. Over a campaign budget, that gap compounds.
The 17% gap on brand equity is more concerning because brand effects accumulate over time. A campaign that builds 17% less brand equity per exposure is not just a one-time loss. It is a structural disadvantage that compounds across campaigns, quarters, and years.
What the Study Actually Measured (and What It Did Not)
The study has important limitations that the headline does not capture:
What It Measured
- Video ads only: Ten pairs of video ads in specific product categories
- Strategy-controlled design: Both AI and human ads were created from the same creative brief, meaning AI was graded on execution, not strategy
- Consumer response: Approximately 3,000 consumers assessed the ads on sales potential and brand equity through Ipsos's validated testing methodology
What It Did Not Measure
- Static ads: The study tested video only. Display, social, and print were not included.
- AI with human oversight: The AI ads were generated without human-in-the-loop refinement. In practice, most teams do not use raw AI output without review.
- Long-term campaign effects: The study measured potential, not actual campaign performance over time.
- All AI models: The study used a specific AI tool. Different models, prompts, and workflows produce different results.
These limitations do not invalidate the findings. They contextualize them. The study tells us something real about raw AI creative output compared to human creative output on the same brief. It does not tell us what happens when a skilled creative director uses AI as a tool rather than a replacement.
The Practical Framework: Where to Use AI in Your Creative Pipeline
The study should not be read as "never use AI for ads." It should be read as "understand where AI adds value and where it does not."
Where AI Strengthens Creative Work
- Concept exploration: Generate dozens of concepts quickly, then have humans refine the strongest
- Production efficiency: Automate repetitive format adaptations, resizing, and localization
- Personalization at scale: Create hundreds of variant messages for different segments, with human oversight on the master creative
- Research and mood boards: Rapid visual exploration to inform human creative direction
Where Human Creativity Remains Essential
- Strategic creative decisions: The brief itself, the core insight, the emotional angle
- Brand voice and tonality: AI tends toward the average, which erodes distinctive brand personality
- Final creative approval: The last review should always involve human judgment, especially for brand campaigns
- Long-form storytelling: The narrative arc that builds brand equity over time
The 17% brand equity gap is the warning here. Brand equity is built through distinctive, memorable, emotionally resonant creative. AI-generated creative, by its nature, tends toward the competent middle. It is good enough to approve but not good enough to build lasting brand value.
What This Means for Small and Mid-Sized Businesses
Small businesses face a specific version of this decision. They often lack the budget for full creative teams and are the most tempted to use AI-generated creative as a complete replacement. The study suggests this is exactly wrong.
The 14% sales potential gap means that AI-only creative is not actually saving money. It is costing you 14% of your campaign's potential sales lift on every campaign. For a small business where every dollar of ad spend matters, that gap is significant.
A better approach for SMBs is to use AI to stretch a limited creative budget, not to replace it:
- Use AI to generate initial concepts and mood boards
- Have a human (even if it is the founder or a freelancer) make the strategic creative decision
- Use AI for format adaptation and personalization
- Keep human oversight on anything that represents your brand to the public
The Bottom Line
The HBR research does not say AI is bad for advertising. It says that substituting AI for human creative judgment, without oversight, produces work that looks competent while underperforming on the business metrics that matter. The gap is real, measurable, and compounding. The companies that will benefit from AI in advertising are those that use it as a tool to amplify human creativity, not as a replacement for it.
Read next:
- OpenAI Sponsored Agents: The AI Ads Revolution SMBs Can't Ignore
- Scale and Google Cloud Publish Enterprise AI Agent Deployment Blueprint
- Wiz MCP Agent Integrations: A New AI Security Layer
- CMI 2026: 95% of B2B Marketers Use AI, Only 39% See Results
Not sure where AI fits in your marketing? Contact Vivid for a creative strategy that uses AI where it helps and humans where it matters.
