DoubleVerify today released regional editions of its 2026 Global Insights report, "Media Quality in the Age of AI," reporting that fraud and invalid traffic violation rates fell 41 percent year over year in North America and 45 percent in Europe, the Middle East and Africa, while artificial intelligence bots generated up to ten times more clicks than humans in some unprotected ad campaigns.

The DoubleVerify research, distributed through the company's Global Insights email newsletter on July 29, 2026, and made available for download on the doubleverify.com/global-insights hub, combines proprietary post-bid measurement data from campaigns the company monitored throughout 2025 with a commissioned survey of 22,000 consumers and 2,020 marketing and advertising decision-makers. Sapio Research conducted the interviews in February 2026 across 21 countries for the consumer sample and 25 countries for the marketer sample, according to DoubleVerify.

Two regional editions, North America and Europe, the Middle East and Africa, are currently available for download. Asia Pacific and Latin America editions are scheduled for release in August 2026, and a French-language version of the EMEA report is due in September, according to the report landing page. The company frames the 2026 edition around a single premise: that generative AI is reshaping both the opportunities and the risks facing the digital advertising industry simultaneously, a framing that has become common across the ad verification sector this year.

Fraud rates fall sharply, though gaps between markets remain

Across both regions studied so far, the headline finding is a steep decline in fraud and sophisticated invalid traffic, known in the industry by the shorthand SIVT. In North America, the Fraud/SIVT Violation Rate dropped 41 percent compared with 2024, though the region still posted the highest rate of any market DoubleVerify tracks, at 0.6 percent, driven mainly by the United States. Canada's rate, measured separately, aligned instead with the regional averages recorded in Asia Pacific, EMEA and Latin America.

In EMEA, the improvement was even larger. The Fraud/SIVT Violation Rate fell 45 percent year over year, continuing what DoubleVerify describes as the world's lowest regional rate for the metric. Non-human data center traffic, the most prevalent fraud type in the region, dropped 54 percent. Rates in Italy and Southern Europe fell further still, while the increase in Sub-Saharan Africa was described as the most dramatic movement in the opposite direction; Turkey's own fraud rate rose even as it remained comparatively low in absolute terms.

The gap between protected and unprotected campaigns was substantial in both regions. Unprotected campaigns, meaning those without DoubleVerify's pre-bid or post-bid safeguards, recorded a fraud violation rate of 2.8 percent globally, compared with 0.6 percent for protected campaigns, a difference the report expresses as 384 percent. That comparison echoes the structure of DoubleVerify's own Q1 2026 quarterly benchmarks report, which similarly found fraud and SIVT violations falling 24 percent year over year to 0.5 percent globally in that earlier dataset, alongside a rising out-of-geo violation rate that ran counter to the broader trend.

Device-level fraud patterns diverge by region

Desktop remained the device attracting the most fraud in both regions, though the direction of the improvement differed. In North America, the Fraud/SIVT Violation Rate on desktop stood at 0.9 percent, against 0.8 percent on mobile app and 0.5 percent on mobile web; desktop showed the steepest year-over-year reduction, at 48 percent. In EMEA, the pattern reversed: mobile app recorded the largest improvement, falling 51 percent to reach a 0.4 percent violation rate, while desktop, at 0.8 percent, improved by 44 percent and mobile web, at 0.2 percent, improved by 49 percent.

DoubleVerify's Fraud Lab, while testing a product called DV AI Verification that launched in late 2025, examined how automated bots are shaping click data on the open web. The lab found that in unprotected media, AI bots generated 15 percent of clicks across the campaigns studied, and that in some individual studies, a monitored ad click was ten times more likely to come from a bot than from a human. The report notes that not all such bots and scrapers are fraudulent by design, but that their activity still counts as general invalid traffic, or GIVT, which should be stripped from campaign performance metrics before advertisers draw conclusions about what is and is not working.

That finding lands amid a broader industry conversation about AI-driven traffic distorting measurement. A State of Click Fraud Report published by Lunio in July 2026 found that 75.6 percent of surveyed marketers estimated losing more than 5 percent of their monthly performance budget to invalid traffic, while only 5.3 percent used a dedicated prevention platform. Separate coverage of a LinkedIn bot-click study quoted Lunio chief executive Nick Morley stating that "the era of AI-powered bots has fundamentally changed the fraud landscape," and warning that fabricated clicks were corrupting the training data that algorithmic bidding systems rely on, not merely wasting budget.

Brand suitability improves but AI content complicates the picture

Brand suitability violation rates, which DoubleVerify defines as instances where a monitored ad serves, or is blocked from serving, within content that fails an advertiser's brand suitability settings, moved in the same favorable direction across both regions, though the size of the improvement was more modest than for fraud.

North America's Brand Suitability Violation Rate fell to 4.1 percent, a 10 percent year-over-year reduction and the lowest of any region DoubleVerify tracks; the figure has now declined for three consecutive years, from 5.4 percent in 2023 to 4.6 percent in 2024 and 4.1 percent in 2025. Within North America, the United States recorded a 4 percent violation rate against Canada's 8 percent, though Canada's figure dropped 11 percent year over year compared with a 9 percent reduction in the United States.

EMEA's rate fell only 2 percent, to 6.9 percent, and the region continued to underperform the 4.8 percent global benchmark. The regional average obscured wide variation by market: France's violation rate dropped sharply, to 4 percent, below the global benchmark, while rates in Italy and Turkey rose year over year, and Eastern Europe recorded the highest subregional rate at 9.9 percent, even as that figure itself fell 14 percent from the prior year.

Unprotected campaigns again showed markedly higher violation rates than protected ones. Globally, unprotected campaigns recorded a 13.8 percent brand suitability violation rate against 4.8 percent for protected campaigns, a gap DoubleVerify expresses as 189 percent. The report's Vodafone Germany case study found that after implementing DoubleVerify's Authentic Brand Suitability product, the telecommunications company reduced media waste by 79 percent and its block rate by 72 percent, alongside CPM and quality-adjusted CPM reductions of 4 percent and 10 percent respectively. Similar deployment findings for Vodafone in EMEA news environments appeared in an earlier PPC Land report from July 2025, which found the company had reduced block rates by 41 percent and keyword violations by 48 percent year over year, redirecting the resulting savings into additional purchases of news publisher inventory.

Where "AI slop" concentrates

DoubleVerify's post-bid content classification, applied across the two regions, identified specific categories where what the industry has taken to calling "AI slop" concentrates most heavily. In North America, the highest volume of AI-generated low-quality content appeared in the Violence category, followed by Travel, Vehicle Disasters and Youth Entertainment. In EMEA, the pattern differed: Travel ranked highest, followed by Technology and Computing and Violence. DoubleVerify says its AI SlopStopper tool blocked or monitored more than 500 million impressions across both regions in the first half of 2026 combined.

The company's broader Fraud Lab has previously documented how such content gets produced at scale. A DoubleVerify investigation published in March 2026 and detailed in PPC Land's coverage of the AutoBait network identified more than 200 Made for Advertising domains using large language models and an AI image generator to produce clickbait articles for roughly 2.25 dollars apiece, generating tens of millions of ad impressions before unprotected advertisers unknowingly paid for placement on what appeared to be independent lifestyle blogs. That report was authored by DV Fraud Lab researchers Arik Nagornov, Merav Geles and Lia Bader.

Because the distribution of low-quality AI content across categories is not predictable, according to the new Global Insights report, DoubleVerify recommends advertisers avoid such content universally rather than attempting to exclude only specific topic categories. The company positions its "Low-Quality GenAI" advertiser profile setting as the mechanism for doing so, a category-agnostic approach that mirrors the logic behind IAS's own Low-Quality GenAI Avoidance product, which moved to general availability in June 2026 after opening in beta two months earlier.

Consumer sentiment toward AI splits sharply by region

Perhaps the most striking divergence in the two reports concerns how consumers in each region feel about artificial intelligence itself, separate from any question of ad fraud or content quality.

In North America, only 50 percent of surveyed consumers said AI tools made their online experience better, against a global average of 63 percent; 22 percent said AI made their experience worse, compared with a global figure of 14 percent. DoubleVerify describes North American sentiment as "strikingly different" from the global pattern and notes that this reluctance carries through to how consumers there view AI-generated advertising. In EMEA, sentiment sat closer to, though still below, the global average: 54 percent of consumers said AI improved their online experience, against 17 percent who said it made things worse.

The two regions also diverged on how marketers there feel about generative AI in ad creative production. North American marketers reported far greater reluctance than their global counterparts about using AI to build ad copy and creatives: 63 percent said they were somewhat to very concerned, against a 48 percent global figure. EMEA marketers showed the opposite pattern, expressing less concern than the global average, at 43 percent against 45 percent globally.

Consumer reactions to AI-generated advertising quality followed a broadly consistent logic across both regions, even where overall sentiment diverged. In North America, 33 percent of consumers said they would feel positive toward a brand using AI to create polished, professional-looking ads, but that figure fell to 23 percent for ads that looked low-quality, uncanny or strange, while negative sentiment for such ads rose to 43 percent. In EMEA, the comparable figures were 33 percent positive for polished ads and 24 percent positive for low-quality ones, with negative sentiment reaching 42 percent for the latter.

Ads inside AI chat platforms draw caution

Both reports addressed a format that barely existed when the surveys were fielded in February 2026: advertising embedded directly inside AI chat platforms. DoubleVerify notes that OpenAI began testing ads within ChatGPT in early 2026, a development PPC Land has tracked since the company confirmed its advertising plans on January 16, 2026, and later expanded through a formal pilot launch on February 9. By the time OpenAI opened its self-serve Ads Manager to all United States businesses on May 5, 2026, the format had moved well past the hypothetical stage DoubleVerify's surveyed respondents were reacting to months earlier.

North American marketers proved the most wary of any group in either report about serving ads inside AI platforms, with 63 percent describing themselves as somewhat to very concerned, against a 50 percent global average; EMEA marketers were closer to that global figure, at 47 percent. Consumer reactions in North America followed a similar pattern, with only the single scenario of "the ad is highly relevant to the chat" producing more positive than negative sentiment, at 31 percent positive against 19 percent negative. Ads appearing alongside chats the report characterizes as negative or sad drew 34 percent negative sentiment in North America, the least favorable result recorded for any chat scenario tested.

As for what would make marketers more comfortable serving ads inside AI platforms, the two regions again showed different priorities. North American marketers ranked transparent reporting on where ads are served and reliable measurement and attribution of ad performance as their top two requirements, both at 39 percent. EMEA marketers instead prioritized the ability to review or approve AI-generated creatives, at 38 percent, ahead of data-privacy safeguards.

Viewability and attention move in opposite directions

Viewability, the measure of whether an ad had the technical opportunity to be seen, improved in both regions, though the scale of the gain differed considerably. North America's overall viewable rate rose 4 percent year over year; video viewability in the United States specifically climbed 10 percent to reach 88 percent, while Canada's video rate rose a more modest 4 percent to 84 percent. EMEA video viewability rose only 1 percent regionally, to 84 percent, still trailing the 87 percent global benchmark, though France recorded the sharpest single-market improvement in display viewability, up 8 percent to 67 percent.

Attention, a metric DoubleVerify calculates through its proprietary Attention Index and indexes against a global benchmark of 100, told a more complicated story. EMEA's overall index stood at 109, above the global benchmark and the strongest of the two regions studied so far; North America's index came in at 97, below benchmark. DoubleVerify's own explanation for the shortfall centers not on content or audience attention spans but on media-buying behavior: in "the world's most mature market," the report states, advertisers "buy programmatically at enormous scale, often prioritizing reach and efficiency" rather than optimizing toward the channels and formats shown to capture attention most effectively. Certain North American formats performed particularly poorly against benchmark, including mobile web medium display ads, which registered an attention score of 90.

What the reports leave open

DoubleVerify's own methodology notes caution that all figures in the report represent DV's proprietary post-bid data filtered to the region studied, and that reported numbers "may not represent the full scope of DV's protection," since some violations are prevented before an impression is ever purchased through the company's separate pre-bid technology. The company derives its wasted-spend estimates by multiplying total violations by an assumed cost per mille of 3 dollars, a figure the report applies uniformly across markets despite considerable variation in actual media pricing between, for instance, the United States and other markets in the same region.

Independent scrutiny of ad-verification measurement more broadly has intensified over the past year. The Media Rating Council issued a policy in October 2025 restricting verification vendors from using the term "brand safety" unless their tools examine actual image, video and audio content rather than relying solely on keyword-level analysis at the domain level, a distinction relevant to any reader evaluating whose brand suitability figures to trust and why. Separately, an Adalytics investigation reported in 2025 alleged that DoubleVerify's own fraud protection services regularly missed non-human traffic served from data center IP addresses, a claim that led to litigation between the two companies and drew the attention of a U.S. senator, who contacted federal regulators over the allegations.

Timeline

Summary

Who: DoubleVerify, a publicly traded media measurement and verification company listed on the New York Stock Exchange under the ticker DV, produced the report; the underlying survey data came from 22,000 consumers and 2,020 marketing and advertising decision-makers interviewed by Sapio Research.

What: DoubleVerify released regional editions of its 2026 Global Insights Report, titled "Media Quality in the Age of AI," covering North America and Europe, the Middle East and Africa, with findings on fraud, brand suitability, viewability and attention metrics alongside survey data on how consumers and marketers view generative AI in advertising.

When: The report was distributed via email on July 29, 2026, and the regional editions became available for download today, July 30, 2026; the underlying consumer and marketer surveys were conducted in February 2026, and the measurement data covers campaigns from January 1 to December 31, 2025.

Where: The North America edition covers the United States and Canada; the EMEA edition covers markets including France, Germany, Italy, Spain, Turkey, the United Kingdom, Saudi Arabia and the United Arab Emirates, among others. Asia Pacific and Latin America editions are scheduled for release in August 2026.

Why: The report matters to advertisers, publishers and marketers because it quantifies, with measured data rather than anecdote, how generative AI is simultaneously creating new advertising formats, new categories of low-quality content, and new sources of automated click traffic, all while consumer trust in AI-generated advertising remains sharply divided between regions such as North America and the Middle East.