AI Ads Can Win Attention. But What Happens to Brand Trust?

As generative AI makes advertising faster and cheaper to produce, South Korean consumer research and new AI transparency rules suggest that brands may need to measure authenticity, credibility and trust alongside clicks, impressions and conversions.

Generative AI is making one of advertising’s oldest problems easier to solve: producing more content, more quickly and at lower cost. Marketers can now create synthetic actors, voices, environments and product visuals, while generating multiple versions of an advertisement for different audiences and markets. But as the cost of producing attention falls, another question is becoming harder to answer: what does that attention do to the consumer’s trust in the brand?

The question is becoming particularly relevant in South Korea, where consumers are increasingly familiar with generative AI but remain cautious about its use in advertising. A survey of 328 Korean consumers who had viewed generative AI advertisements found that 71% experienced some degree of aversion to them, while 88% preferred human models over AI-generated virtual models. At the same time, 94% said AI-generated content should be subject to mandatory labeling.

The findings point to a contradiction at the center of AI advertising. Consumers may be able to recognize synthetic content, and many want brands to disclose it, but disclosure and realism do not necessarily translate into greater acceptance.

For brands, that means conventional advertising metrics may no longer tell the whole story.

“Getting attention has never really been the hard part,” Atique Bandukwala, Founder and CEO of Vidopix, told KoreaTechToday. “You can get attention with a loud noise.”

His argument is that the industry has spent years becoming better at measuring whether people watched an advertisement, clicked on it or completed it, without necessarily measuring what the creative did to the relationship between the consumer and the brand. As AI-generated advertising becomes more common, that gap could become a significant business and reputational risk.

AI is changing the economics of advertising

Generative AI is already becoming embedded across advertising workflows, from copywriting and image creation to video production, personalization and campaign testing. A 2026 review published in the International Journal of Advertising describes AI as increasingly integrated into the advertising ecosystem and notes that the technology is reshaping both creative production and consumer interaction with advertising. The appeal for marketers is straightforward.

A campaign that once required photographers, actors, locations, editors and multiple rounds of production can increasingly be generated or modified with software. Brands can create more variations and test them across audiences without incurring the same production costs associated with conventional campaigns. That changes the economics of creative experimentation.

But it also creates a potential measurement problem. Traditional advertising dashboards are built around outcomes such as impressions, clicks, viewing time, completion rates and conversions. These remain useful indicators of campaign performance, but they do not necessarily explain whether an audience found the advertisement credible, whether it strengthened brand associations or whether consumers would react differently after learning that the content was synthetic.

“Most of them are not being thoughtful about it at all,” Bandukwala said of brands’ current approach to AI-generated advertising. “They are measuring clicks and impressions and completion rates, which tells you whether someone watched the ad but tells you absolutely nothing about whether they trusted it or whether it moved the needle on how they feel about the brand.”

That distinction could become more important as AI makes it possible to optimize creative for attention at unprecedented scale.

Realistic does not necessarily mean authentic

One of the assumptions behind synthetic advertising is that better technology will make consumers more accepting. If an AI-generated face looks human enough, a synthetic voice sounds natural enough and a generated environment is convincing enough, the thinking goes, the audience may simply respond to the advertisement as it would to conventional creative. Research increasingly suggests the relationship is more complicated.

A 2025 study published in Management Research Review examined consumer evaluations of AI-generated advertising through the concept of perceived authenticity. It found that the question is not simply whether AI can reproduce human creative elements, but how consumers interpret the role of AI in producing those elements.

More recent research has similarly identified competing effects in AI-generated advertising, including credibility on one side and perceptions of eeriness on the other. A 2026 study in Acta Psychologica argues that AI advertising can activate these competing pathways, meaning that technological realism alone does not determine whether consumers trust the advertisement. This helps explain the Korean consumer data.

Two-thirds of respondents in the 328-person survey said they could distinguish AI-generated advertisements from human-created ones. Yet 71% still reported feeling aversion toward AI-generated advertising, and 88% preferred human models.

Korea’s consumers are asking for transparency

South Korea offers an unusually useful environment for examining this issue because consumer expectations and regulation are moving in the same direction. The Korean survey found that 94% of respondents supported mandatory labeling of AI-generated content. The survey was not a nationally representative population study, so the result should not be interpreted as a definitive measure of all Korean consumers. But it does provide a strong signal that transparency is becoming an important part of the consumer conversation around synthetic advertising.

The country’s regulatory framework is also moving toward greater transparency. South Korea’s AI Basic Act establishes transparency obligations for generative AI. Article 31 requires AI businesses providing generative AI products or services to notify users in advance and identify outputs as being generated by generative AI. For AI-generated audio, images or video that are difficult to distinguish from reality, the law requires disclosure or labeling in a way that users can clearly recognize. The relevant provisions took effect on July 21, 2026.

That changes the role of disclosure. It is no longer simply a question of whether a marketing team thinks telling consumers about AI use is strategically advantageous. For certain types of synthetic content, transparency is becoming part of the operating environment in which brands have to work. But transparency introduces another complication.

The transparency paradox

One might assume that simply labeling an advertisement as AI-generated would solve the trust problem. Research suggests it is not that simple. A 2026 Korean study examining consumer attitudes toward AI-generated advertising found that consumers generally evaluated traditional advertisements more favorably than AI-generated advertisements. It also found that AI literacy influenced how consumers responded, with people who had greater familiarity with AI tending to evaluate AI-generated advertisements more positively and showing less sensitivity to disclosure.

Other recent research points to an even more complicated relationship between disclosure and trust. A 2025 study of generative AI in service advertising found that AI disclosure could reduce trust and produce less positive advertising attitudes in certain contexts. That does not mean brands should conceal AI use. In an environment where transparency requirements are becoming stronger, doing so could create an even greater credibility and compliance problem.

Instead, it suggests that disclosure is only one part of the trust equation. Consumers may accept AI when they understand why it is being used, when the use fits the brand and when the resulting creative still feels credible. They may react differently when AI appears to be replacing human authenticity or creating an impression that the brand is manufacturing something that consumers are expected to believe is real.

A recent 2026 study offers another piece of the puzzle. Researchers found that verification mechanisms such as QR codes and third-party certification signals could increase consumer trust in AI-generated advertising, although their effects on purchase intention differed. The emerging lesson is therefore not that AI advertising is inherently untrustworthy. It is that trust requires signals beyond the advertisement itself.

The metric problem for brands

This is where the advertising industry’s measurement model may need to evolve. A high click-through rate can indicate that an advertisement generated interest. A high completion rate can indicate that viewers stayed until the end. A strong conversion rate can show that a campaign contributed to a transaction. But none of these metrics necessarily tells a brand whether consumers trusted the message.

That becomes particularly important for synthetic advertising because the same creative can produce two very different outcomes. A realistic AI-generated spokesperson might attract attention because the content looks novel. But if consumers later discover that the person does not exist, the novelty could become a credibility problem.

A synthetic environment could make a campaign visually distinctive. But if it creates an uncanny or artificial feeling that conflicts with the brand’s identity, the advertisement could strengthen short-term engagement while weakening longer-term brand perception. This is why Bandukwala argues that brands need to expand the definition of advertising performance.

“What brands should be looking at is emotional resonance, brand recall, sentiment shifts over time, and whether the creative actually aligns with the advertising standards that apply in each market,” he said.

These are harder variables to measure than clicks. They are also closer to what brands ultimately want advertising to accomplish.

The distinction can be summarized simply:

The point is not to replace performance metrics with softer measures of sentiment. It is to recognize that immediate performance and long-term brand value are not always the same thing.

AI may make pre-publish testing more important

Generative AI also changes the point at which brands can evaluate creative. Historically, much advertising optimization happened after launch. Brands could analyze audience behavior, identify drop-off points and study sentiment after the campaign had already spent part of its media budget.

AI can now be used to test creative before publication. Vidopix, which Bandukwala founded, is building its business around this idea. The company describes its platform as a pre-publish video intelligence system that analyzes attention, emotion, brand recall and compliance before media spending begins. Its Pixi system is designed to provide frame-level analysis and identify creative risks before an advertisement goes live.

The company says its platform has analyzed more than 450,000 ad creatives, tracks more than 27 emotion dimensions and supports more than 250 languages. Those are company-reported figures rather than independently audited market data.

The broader idea, however, extends beyond Vidopix. If brands can generate hundreds of AI-assisted creative variations, they will need ways to determine which ones are likely to produce the desired response before deploying them at scale. That makes pre-publish testing increasingly important. But the definition of testing may also need to expand.

It is not enough to ask whether an AI-generated advertisement will hold attention. Brands may also need to ask whether it creates discomfort, whether the AI element distracts from the brand, whether the creative produces the intended emotional response and whether the use of synthetic content introduces a regulatory or cultural risk.

Bandukwala’s own view reflects this broader shift. At Vidopix, he said, the company’s Pixi platform includes a compliance layer that allows brands to upload their own regulatory frameworks and flag rule-level issues before content is published.

“That last part is becoming increasingly important because regulators around the world are starting to move on this, and they are moving faster than most marketing teams seem to realise,” he told KoreaTechToday.

For brands operating across Asia, that could become a significant challenge. An advertisement developed for one market may encounter different disclosure requirements, advertising standards, cultural expectations or rules around synthetic media in another.

Korea could become an important test market

South Korea is particularly interesting because it combines high digital adoption with rapid AI commercialization and increasingly explicit AI governance. A 2026 Digital Lifestyle Report from CJ Mezzomedia found that 88% of surveyed Korean consumers had experience using generative AI services. The report also examined consumer perceptions of generative AI advertising, highlighting how quickly AI has moved from an emerging technology into everyday digital behavior.

That creates a market in which consumers are likely to encounter increasingly sophisticated synthetic advertising while also becoming more familiar with the technology behind it.

For advertisers, that creates a moving target. Consumers may become better at recognizing AI-generated content as the technology improves. At the same time, familiarity with AI could make some audiences more accepting of it. The 2026 Korean academic research on AI literacy points in this direction: consumers with greater AI literacy were more favorable toward AI-generated advertising and less sensitive to disclosure. This suggests that the future of AI advertising will not be determined by a simple question of whether consumers like or dislike AI. Acceptance is likely to depend on who the consumer is, what the brand represents, how AI is used, how transparent the advertiser is and what the creative is trying to communicate. That makes audience-level measurement more important, not less.

What brands should measure beyond clicks

For marketers, the emerging framework is not about abandoning conventional performance analytics. It is about adding measures that capture the effects AI can introduce.

The most important areas include:

• Emotional resonance: Does the creative generate the intended emotion, or does the synthetic element create discomfort or distance?
• Brand recall: Do consumers remember the brand rather than simply remembering the AI-generated character or visual?
• Sentiment over time: Does consumer sentiment remain positive after people learn that AI was involved?
• Perceived authenticity: Does the creative feel consistent with the brand’s identity and values?
• Transparency and compliance: Is AI use disclosed appropriately for the relevant market, platform and content type?
• Long-term brand impact: Does the campaign strengthen or weaken trust beyond immediate engagement?

These measures will not replace impressions, clicks or conversions. Instead, they can provide context around them. An advertisement that generates high engagement and strong trust is strategically different from one that generates high engagement but leaves consumers feeling deceived or disconnected from the brand. That distinction could become increasingly important as AI lowers the cost of producing creative.