In performance advertising, generative AI has moved from an experimental tool to a core capability. In the past, creative production took days or weeks, but it can now happen in minutes at a scale previously unheard of.

However, as consumers see more and more AI-generated content, a critical question is emerging: Can you build the same level of trust and emotional connection with machine-generated creative as with work shaped by human hands?

Recent research suggests that this question may be framed incorrectly.

The real issue is not whether AI can perform, but whether audiences perceive AI-generated creative as authentic. When ads feel artificial or overly synthetic, engagement declines. When AI is used to amplify human-centric creative principles, however, performance does not merely hold steady — it can improve. As it turns out, trust is not a soft brand value in AI-driven campaigns, it’s a measurable performance constraint.

The Evolution of AI in Performance Advertising

The role of AI in performance advertising has expanded in stages, moving from backend efficiency tools to a highly visible force in creative production. Each phase has increased AI’s impact on results, while also raising new questions about how audiences perceive and respond to machine-generated content.

Here’s a quick look at the evolution in performance advertising:

Stage 1: Efficiency & Optimization

Initially, AI operated almost entirely behind the scenes. It was primarily used to manage bids, segment audiences, and allocate budgets in real time. These systems improved campaign efficiency and reduced marketers’ manual workload, but they remained invisible to consumers.

Because AI did not directly shape what audiences saw, there was little risk to brand trust or creative authenticity.

Stage 2: Content Proliferation

As generative models matured, AI moved closer to the creative layer. Tools such as Realize’s GenAI Ad Maker enabled the generation of thousands of visual and textual variations at virtually no additional cost. This shift democratized high-volume creative testing, allowing advertisers to experiment at a scale that was previously reserved for the largest budgets. For performance marketers, AI became a powerful engine for rapid iteration and optimization, significantly accelerating learning cycles.

Stage 3: The Authenticity Era (Current)

Today, AI sits at the forefront of creative production, generating assets that consumers actively see and judge. As a result, research attention has shifted toward “perceived artificiality” and its effect on engagement.

The industry is increasingly recognizing that while AI can calculate, optimize, and scale with remarkable efficiency, emotional connection still depends on human-centric creative cues. In this phase, authenticity is no longer a creative preference but a performance requirement.

Key Findings: Why Human Connection Wins

Large-scale academic research has begun to clarify why human connection remains central to the success of AI-driven advertising. In a recent study led by teams from Columbia University, Harvard University, the Technical University of Munich, and Carnegie Mellon, researchers analyzed live performance data for ads created with Realize’s GenAI Ad Maker. Working in collaboration with Taboola’s in-house creative agency, Creative Shop, the research examined how AI-generated and human-made ads perform side by side across real advertisers, real audiences, and real consumer behavior.

Instead of relying on surveys or controlled lab simulations, the study drew on in-market campaign data. This allowed researchers to observe how perceptions of authenticity influence engagement at scale. This real-world approach makes the findings particularly relevant for performance marketers, where creative decisions are ultimately judged by how audiences actually respond, not how they say they might.

Here is a summary of some key findings:

1. The “AI Penalty” vs. the Disguise Advantage

One consistent finding across the research is the presence of what behavioral scientists describe as “algorithm aversion.” Consumers often bring a negative predisposition toward content they believe was generated by a machine, particularly when it appears overly polished or formulaic. This reaction can suppress engagement even when the underlying message is relevant.

Overall, AI-generated ads tend to perform on par with human-made ads in terms of click-through rate (CTR). That parity alone is notable given the speed and efficiency advantages AI provides. However, averages obscure an important pattern: When researchers examined performance through the lens of perception, a clear hierarchy emerged. AI-generated ads that were not perceived as artificial achieved the highest engagement rates, outperforming both obviously AI-made ads and traditional human-made creative.

Side-by-side comparisons from live campaigns show how closely AI-generated and human-made ads can perform when they follow the same creative principles. In fact, in the following matched examples (see image below), the AI-generated ads achieved slightly higher CTRs without appearing artificial to viewers.

The Columbia study backs this up. Across the dataset, which included over 300,000 ads, human-generated ads had a CTR of around 0.65%, while AI-generated ads had a CTR of around 0.76%.

The inverse was also true. Ads that felt synthetic were penalized by audiences, regardless of whether they were actually created by AI or by humans. In practical terms, this means performance is governed less by how an ad is made and more by how it is perceived. AI’s advantage lies in its ability to convincingly blend into the visual and emotional language audiences associate with human communication.

2. Faces as the Bridge to Trust

Among all the variables examined in recent advertising research, few are as consistently powerful as the presence of human faces. From an evolutionary perspective, humans are biologically wired to quickly notice faces and draw emotional information from them. Faces signal intention, relatability, and social presence, all of which help reduce uncertainty in fast-moving digital environments.

In performance advertising, this biological bias becomes a functional advantage. Ads featuring clear, prominent human faces are more likely to be perceived as authentic and human-made, which in turn supports stronger engagement. Research examining visual attributes confirms that facial presence is one of the strongest predictors of whether an ad feels trustworthy.

Global clothing brand H&M understands this. In 2025, they replaced some of their fashion models with AI-generated versions. However, to ensure the digital models “reflected the real model’s individuality,” they worked with AI specialists and creative teams to design the AI models and build the campaign narrative. This was done to ensure that the technology wasn’t taking away from human artistry, but building upon it.

Interestingly, AI-generated creative often includes human faces more frequently than human-designed ads. This is not an accident: AI systems trained on large volumes of high-performing creative tend to replicate the patterns that already work, including the consistent use of faces to establish emotional connection. When applied thoughtfully, this can help AI-generated ads overcome skepticism rather than amplify it.

3. Visual Cues of Artificiality

While faces help bridge trust gaps, certain visual characteristics reliably signal artificiality and trigger disengagement. Ads that rely on extreme sharpness, heavy color saturation, or highly symmetrical compositions are more likely to be perceived as machine-generated. These traits can create a sense of visual perfection that feels detached from real-world experience.

By contrast, ads that incorporate warmth, natural composition, and subtle imperfection tend to feel more human. High realism, rather than hyper-polish, supports relatability. In this context, the goal is not to disguise AI through deception, but to align creative output with the visual cues audiences already associate with authenticity. Avoiding these perceptual markers is often enough to preserve trust and maintain performance.

Key Takeaways: Trust as a Performance Metric

Generative AI is reshaping how performance advertising creative is produced, but it has not rewritten the fundamentals of human psychology. Audiences still respond to authenticity, familiarity, and emotional clarity. The results of the Columbia study, along with other research, show that AI can support these outcomes rather than undermine them, provided it is guided by human-centric design.

For advertisers, this reframes the AI debate. The choice is not between human creativity and machine efficiency, but between scaling distrust or scaling connection. Authentic faces, relatable contexts, and emotionally grounded visuals are no longer optional refinements: They are prerequisites for sustainable performance.

Frequently Asked Questions (FAQs)

Does using AI in ad creative damage consumer trust?

Consumer skepticism toward AI-generated content is real, but it is not absolute. Research shows that distrust tends to arise when ad creative feels robotic, overly perfect, or detached from human experience. According to Frontiers in Psychology, in some functional contexts, such as data-driven messaging or informational content, transparency about AI use can even reinforce perceptions of competence. Problems tend to emerge when emotionally or narrative-driven creative lacks human nuance, resulting in what many marketers now refer to as “AI slop.”

In other words, telling people that content was created by AI changes how they react to it. When the content is practical or informational, being upfront about AI can actually help, as people see it as efficient and objective. In fact, a 2024 Sprout Social survey found that 94% of consumers believe all AI content should be disclosed. But, when the content is meant to entertain, inspire, or connect emotionally, revealing that it was made with AI can backfire, making it feel less genuine or engaging.

So, knowing when and when not to disclose is critical, but you can also mitigate AI distrust with Realize by using its GenAI Ad Maker tool to prioritize human-centric signals. Research that analyzed large volumes of live ads found that a significant share of AI-generated creative was perceived as human-made, suggesting that thoughtful design can neutralize the so-called AI penalty. By emphasizing elements like authentic faces and natural composition, AI-generated ads can blend seamlessly into human content ecosystems.

How can AI-generated ads achieve a “human touch”?

You need more than prompt engineering to humanize AI-generated creative. Try removing repetitive phrasing, focus on real audience pain points, and allow room for humor or narrative context. In practice, the most successful teams treat AI as a collaborator rather than a replacement, using human judgment to guide tone and emotional direction.

Realize embeds this humanization directly into its AI systems by training models on proven creative best practices. When AI is informed by what already resonates across large publisher networks, it can replicate human-like patterns at scale. This includes automatically prioritizing visual elements that audiences associate with authenticity, reducing the need for constant manual oversight.

What are the performance benefits of using AI for human-centric ads?

AI’s primary advantage remains its ability to personalize and iterate at speed. High-velocity testing allows advertisers to identify effective messages and visuals faster than manual workflows permit, improving engagement and return on investment. Importantly, research examining downstream performance suggests that increases in click-through rate driven by AI-generated creative do not come at the expense of conversion quality.

On the Realize platform, AI-generated ads that adhere to human-centric principles can outperform traditional creative while maintaining conversion integrity. This combination of scale, efficiency, and trust preservation represents the most compelling use case for AI in performance advertising today.