DoubleVerify finds AI slop is becoming a brand safety problem
DoubleVerify data connects AI slop with weaker brand perception, distorted click metrics, and higher media quality demands.
DoubleVerify has put a sharper definition around a problem advertisers have mostly treated as an aesthetic nuisance: low-quality AI-generated content is becoming a media quality risk.
Its latest Asia Pacific findings connect three issues that often sit in separate conversations. Brands can appear beside synthetic content that weakens perception, publish AI-assisted creative that audiences judge as cheap or unnatural, and optimize against clicks that may not come from people at all. Together, those risks make AI slop less a content problem than a signal integrity problem.
Table of contents
Jump to each section:
- Why AI slop changes the media quality equation
- Quality is becoming a placement signal
- AI chat ads face a different trust test
- What marketers should know about AI slop
Why AI slop changes the media quality equation
Brand safety has traditionally focused on avoiding obviously harmful or unsuitable material. AI-generated content complicates that model because the risk can be softer and more cumulative. A page does not need to contain hate speech or misinformation to make a placement feel careless. It may simply be repetitive, spam-like, poorly rendered, or engineered to attract automated traffic.
49% of Singapore consumers said seeing an ad beside low-quality or spam-like AI-generated content would negatively affect their perception of the brand.
The placement still works in a technical sense. The ad loads, an impression is counted, and the surrounding content may pass conventional exclusions. But the environment can quietly transfer its low-effort character to the advertiser.
Media quality is no longer only about what surrounds an ad. It is also about what the surrounding experience signals about the brand's judgment.
The common assumption is that adjacency risk can be handled with blocklists and category controls. The contrasting reality is that AI slop often looks acceptable at the category level while failing at the quality level. The strategic implication is that suitability systems will need to evaluate production patterns, page intent, and traffic quality, not just subject matter.
Patrecia Meliana
Quality is becoming a placement signal
The same quality judgment applies to the ad itself. The findings suggest audiences are not rejecting AI creative as a category. They are distinguishing between work that feels deliberate and work that exposes the shortcuts behind it.
47% of Singapore consumers said low-quality, uncanny, or unnatural AI-generated advertising would negatively affect their view of a brand.
That distinction matters because it weakens a convenient defense of poor AI creative. A negative response cannot simply be dismissed as resistance to new technology. Consumers appear capable of rewarding polished execution while penalizing output that feels unfinished.
AI does not remove the cost of craft. It moves that cost from production into selection, direction, and quality control.
For creative teams, faster generation can produce more options, but more options also create a harder editorial burden. Someone still needs to decide which asset carries the right visual logic, emotional tone, and level of finish for the context. Scale without that judgment can increase the volume of acceptable assets while reducing the share that feels distinctive.
The measurement problem is even less visible. Automated agents, crawlers, and other nonhuman traffic can interact with ads in ways that resemble engagement, creating signals that optimization systems may reward.
15% of clicks in unprotected media were attributed to AI bots in DoubleVerify's testing, showing how automated activity can contaminate campaign metrics.
A campaign can therefore look efficient for two bad reasons at once: inexpensive inventory and artificial interaction. That combination is dangerous because it turns low quality into a positive performance signal. Once an optimization system learns from distorted clicks, the budget can be pushed toward more of the same inventory.
AI chat ads face a different trust test
Conversational platforms add another layer. In a feed or on a publisher page, an ad is visibly separate from the content around it. Inside an AI exchange, the advertising experience may sit closer to the user's question, recommendation process, or purchase decision.
42% of Asia Pacific consumers said a closely relevant ad within an AI conversation would positively influence their perception of the advertised brand.
Relevance is promising, but it also raises the standard. A conversational ad that is contextually useful can feel like assistance. One that is poorly matched can feel more intrusive than a conventional display placement because it interrupts an exchange that users experience as personal and purposeful.
The more interesting question is not whether AI platforms can carry ads. It is whether advertisers can verify the conditions that make those ads trustworthy.
Marketers will need evidence about where the placement appeared, how the platform distinguished sponsorship from an organic answer, what interaction was measured, and whether the response came from a person. These are not separate concerns. Creative control, suitability, disclosure, and measurement converge inside the same interface.
What marketers should know about AI slop
The practical lesson is to treat AI content quality and traffic quality as one connected media discipline.
Expand suitability beyond sensitive topics. Review whether verification settings can identify repetitive, spam-like, or low-value synthetic environments. Category exclusions alone may not capture the reputational risk.
Separate generation speed from approval speed. AI can accelerate asset creation without shortening the judgment required before release. Creative review should test for finish, naturalness, brand distinctiveness, and contextual fit.
Audit the signals feeding optimization. Clicks and engagement should be filtered for automated activity before they influence bidding, creative rotation, or attribution. A clean dashboard is not useful if the underlying audience is unclear.
Set a higher evidence bar for conversational ads. Before moving budget into AI chat environments, ask how relevance, disclosure, placement quality, and human interaction are verified. The interface may be new, but the accountability still belongs to the advertiser.
AI slop reveals a broader shift in digital advertising. Supply is becoming easier to manufacture, whether that supply is content, creative, inventory, or engagement. As abundance rises, the scarce capability is not production. It is knowing which signals deserve trust.
That changes the role of media quality from a defensive control into a strategic input. The teams that can distinguish real attention from synthetic activity, and deliberate creative from automated filler, will be better positioned to use AI without allowing its weakest outputs to define the brand.
