For chief marketing officers, chief financial officers, and performance marketing leaders, the question surrounding generative artificial intelligence is no longer whether it works, but whether it improves return on investment. Creative production has become one of the most costly parts of digital advertising, and performance expectations continue to rise: Advertisers are under pressure to test more variations, personalize at greater scale, and maintain efficiency in increasingly competitive auctions.

Against this backdrop, artificial intelligence (AI)-powered ad creative has emerged as a potential solution, but it brings perceived risk — namely, that if AI-generated ads cause skepticism or disappointment, any efficiency gains will quickly disappear.

That challenge is exactly what a recent large-scale study set out to explore. Conducted by researchers from Columbia Business School, Harvard Business School, the Technical University of Munich, and Carnegie Mellon University, the study examined whether AI-generated ads — created by Realize’s GenAI Ad Maker — could deliver results without increasing costs or sacrificing quality.

Their findings challenge several assumptions that have shaped how leaders think about AI creative.

Debunking the Curiosity Myth

An ongoing concern with AI-generated ads is the idea of “curiosity clicks.” The worry is that ads created by artificial intelligence may attract attention simply because they stand out as unusual or novel, rather than because they resonate with genuine user intent. Over time, this type of engagement could undermine not only ad performance, but overall brand trust.

The research disputes this.

What matters most, per the data, is perceived artificiality, not how an ad was actually made. Researchers describe this reaction as algorithm aversion, which is a negative predisposition toward content consumers believe was machine-made. Ads that look synthetic are penalized, even when they’re created by humans. On the other hand, AI-generated ads that convincingly blend in with surrounding content avoid this penalty entirely.

Those AI-created ads didn’t just go undetected alongside the human-made ones, though: The study found that, when indistinguishable from their human-made counterparts, these ads achieved what they called “superhuman” click-through rates (CTRs), exceeding traditional ads.

This is where cost-efficiency enters the picture. Tools such as GenAI Ad Maker allow advertisers to scale creative output rapidly, but scale alone is not the advantage: The real advantage is that quality and authenticity can be preserved even as volume increases, as long as the right creative conditions are met.

Human Cues: The Face of Success

Among all the visual characteristics examined, one factor stood out above the rest: Ads that featured large, clear human faces were far more likely to be perceived as human-made and, as a result, achieve higher engagement.

The finding reinforces what performance advertisers have seen for years across eye-tracking studies, platform benchmarks, and creative effectiveness reports. Human faces play a critical role in fast-moving, feed-based environments because they:

• Capture attention quickly.
• Convey emotion and relatability at a glance.
• Establish trust in crowded content streams.

What surprised the researchers was how consistently AI-generated ads applied this principle. In the dataset, creatives produced with GenAI Ad Maker were more likely to include prominent human faces than traditional human-made ads. That emphasis helped AI-generated ads blend naturally into surrounding content. It also explained why many of these ads matched or exceeded human performance when they avoided overly stylized or artificial aesthetics.

The takeaway isn’t that AI changes the rules of effective ad creative. Instead, it shows that AI can apply longstanding, human-centered best practices reliably and at scale, reinforcing performance instead of undermining it.

Matching Human Results at Scale

The conclusions in this study aren’t based on small A/B tests or limited pilots. Rather, they come from real-world performance data analyzed at significant scale, offering a rare view into how AI-generated creative performs in live campaigns.

Across the full dataset, the results were clear:

• 300,000+ live ads were analyzed across multiple industries.
• AI-generated creatives averaged a 0.76% CTR.
• Human-made ads averaged a 0.65% CTR.

These numbers suggest a performance advantage for AI-generated creative. The research went a step further, though, to ensure those results weren’t influenced by external factors such as targeting differences, timing effects, or campaign structure.

To accomplish that, the researchers applied a “sibling ads” methodology. In this approach, AI-generated and human-made ads were matched so that they:

• Came from the same advertiser.
• Ran in the same campaign.
• Launched on the same day.
• Shared the same objective and landing page.

Under these tightly controlled conditions, AI-generated and human-made ads showed a performance that was statistically comparable.

For decision-makers, this distinction is important. Matching human performance under strict controls shows that AI-generated ads aren’t simply benefiting from looser conditions or novelty effects. Instead, they can deliver human-level outcomes reliably and at scale, creating room for efficiency gains without risking performance.

Efficiency for the CFO: High Performance, Low Cost

From a financial standpoint, the promise of AI-powered creative is no small improvement: It signals a substantial shift in the way ad spend is allocated. Traditional creative production relies on human labor, agency workloads, and fixed costs that scale poorly. A recent survey of marketing leaders found that 76% are investing in generative AI solutions, with more than half specifically targeting creative production to cut costs and speed up workflows.

AI tools such as GenAI Ad Maker simplify creative generation. They enable rapid creation of ad variations at a fraction of the cost and time associated with manual production. Most importantly, the study found no evidence that higher AI-driven CTRs came at the expense of conversions: performance held steady even as creative volume increased.

For chief financial officers (CFOs) focused on efficiency and risk management, the combination of stable conversion outcomes while cutting costs is appealing. It allows teams to shift resources away from production and toward testing and improving ad performance.

5 Ways AI-Creative Campaigns Offer Scalable Results at a Lower Cost

Cost efficiency in advertising isn’t about cutting corners, it’s about removing friction from the parts of the workflow that can slow learning, limit testing, and consume budget without improving outcomes. Below are five ways AI-driven creative campaigns help advertisers achieve stronger performance at lower cost, without sacrificing quality or control.

1. Faster Time to Learning

In performance marketing, speed to insight is often more valuable than perfection. Multiple industry reports show that creative is one of the biggest bottlenecks in optimization cycles. This isn’t because teams lack ideas, it’s simply because producing, testing, and deploying new variations takes time.

Unfortunately, regular testing is still crucial to success. Studies find that increasing the velocity of creative testing leads to better performance outcomes, with brands that A/B test ad creative on a weekly basis seeing 31% higher conversion rates than those that test less frequently.

AI-powered creative workflows compress those learning loops. By generating and launching multiple creative variations quickly, advertisers can gather useful insights early in a campaign’s lifecycle. That feedback helps teams refine messaging, visuals, and formats while budgets are still flexible, rather than discovering issues after the money has already been spent.

2. Lower Marginal Cost per Variant

With traditional creative production, scaling comes at a cost. Each new asset requires additional design time, revisions, approvals, and coordination across internal or external teams. For many organizations, that means the marginal cost of testing “one more idea” is high enough to limit experimentation. Organizations cite creative production as one of the most resource-intensive aspects of modern digital marketing, particularly as teams attempt to deliver personalized and multi-format campaigns at scale.

AI shifts that equation by making it easy to generate additional assets without requiring more resources. This allows advertisers to expand coverage without increasing headcount or spend, improving cost efficiency while also encouraging experimentation.

3. Reduced Creative Fatigue Risk

Creative fatigue shows up in a simple but costly way: people tune out. When audiences are repeatedly served the same visual and message structure, engagement drops because the ad no longer feels like fresh information. One industry report found that 93% of customers skip or block ads. In crowded, feed-based environments, that means advertisers are often paying for impressions that don’t drive results.

A few things typically contribute to creative fatigue:

• Repeated exposure to the same visuals.
• Messaging that becomes predictable over time.
• Lack of variation across formats or placements.

AI-driven creative helps reduce this fatigue by making refresh less of a burden. Instead of restarting the production process for every update, teams can generate and rotate new variations frequently while preserving the same offer and core message. As a result, brands can keep things fresh without constant manual effort.

4. Consistent Application of Proven Creative Signals

Long before generative AI entered the conversation, advertising research consistently pointed to the same conclusion: Human cues drive attention. Clear focal points, realistic imagery, emotional relevance, and human presence get results. That includes the use of faces in ads, which studies have shown perform better than other visual stimuli.

Key findings from a recent study reveal:

• 91.7% of ads featuring a human face attracted more attention than ads without faces.
• Faces are detected at least twice as fast as many other visual elements.
• When models make direct eye contact, viewers are more likely to perceive trust and emotional connection.
• Viewers subconsciously follow the direction a person in an ad is looking.

AI systems trained on large datasets apply these principles consistently across hundreds or thousands of assets. Instead of relying on individual designers to interpret best practices asset by asset, AI ensures human-centric cues are repeated reliably at scale. That consistency helps protect performance as volume increases, which is something that’s difficult to maintain manually.

5. Budget Reallocation Toward Media Performance

The shift toward AI doesn’t just reduce production costs, it frees up funds that businesses can shift toward other efforts. When creative absorbs a large share of the marketing budget, it constrains how much can be invested in reach, testing, and optimization. Since marketing leaders are under increasing pressure to demonstrate short-term return on investment, it’s important to assess how much spend is tied up in fixed operational costs rather than flexible, performance-driven investments.

Lowering creative overhead lets organizations redirect spend toward the levers that most directly influence outcomes, including:

• Expanding reach into new or under-tested audiences.
• Increasing testing velocity across creatives, formats, and placements.
• Funding faster optimization through bidding and budget adjustments.
• Extending the lifespan of high-performing campaigns without creative bottlenecks.

By reducing the fixed cost of creative production, AI enables this shift without forcing trade-offs in quality or volume. For business leaders, the result is improved capital performance. The same overall budget delivers more testing, more reach, and more opportunities to optimize.

Key Takeaways

Generative AI is changing the economics of performance marketing. When applied strategically, AI-driven creative enables faster experimentation, broader coverage, and sizable cost efficiencies without sacrificing effectiveness. At scale, AI-generated ads can perform on par with human-made creative while removing many of the operational constraints that traditionally limit testing. The advantage doesn’t come from automation for its own sake: Results improve when AI is used to reinforce established creative best practices, producing ads that feel authentic, credible, and designed for how people actually engage.

Frequently Asked Questions (FAQs)

Do AI ads drive accidental or low-quality clicks?

Some advertisers worry that AI-generated ads may attract attention because they feel unfamiliar or “uncanny,” leading to clicks driven by curiosity rather than genuine interest. In real-world campaigns analyzed in the study, though, higher engagement from AI-generated creative (in this case, from GenAI Ad Maker) did not come at the expense of outcomes. The study found no evidence that increased click-through rates led to weaker conversion performance, indicating that AI-driven engagement reflected real intent rather than accidental interaction.

What visual “red flags” make an ad look like AI to consumers?

Audiences tend to disengage from creative that feels overly artificial or synthetic. When visuals appear too perfect or unnatural, viewers may subconsciously distrust the content, leading to lower engagement and weaker brand perception. Ads that trigger that reaction often feature exaggerated polish, odd proportions, or human elements that feel subtly “off.”

Insights from campaigns analyzed in the study highlight specific design patterns that increase perceived artificiality. Heavy color saturation and highly stylized, glossy visuals were most likely to signal AI-generated content. In contrast, ads featuring large, clear human faces consistently appeared more authentic to viewers and earned stronger click-through performance.

What visual features make an ad look too much like AI?

When creative leans too far into artificial-looking elements, audiences may instinctively pull back. Content that feels synthetic or exaggerated can trigger skepticism, reducing trust and engagement even before the message is processed. Analysis from campaigns found that excessive color saturation and overly polished design were the strongest indicators that an ad would be perceived as AI-generated. Creative that emphasized larger, more natural-looking faces and simpler, less saturated visuals was far more likely to be seen as authentic.