Artificial intelligence is changing how people, platforms, and machines interact with the web. In a recent post, we talked about the growing noise around “AI slop” and the importance of clearer thinking and better definitions when it comes to monetizing synthetic content. That same mindset applies to the current conversation around AI-driven media traffic.
The 2026 IAB Outlook Study highlights that “Adapting to changing consumer habits is now the top investment challenge (44%), surpassing macroeconomic factors, while 38% of buyers cite understanding generative AI as a major challenge, up 14 points from 2024, underscoring the need for deeper AI fluency and strategic application.”
The question advertisers should be asking is not whether AI exists in the ecosystem. It is whether we are measuring it accurately and understanding its real impact.
Automation ≠ Fraud
Industry headlines increasingly suggest that a growing share of ad clicks now comes from “AI bots.” That framing sounds straightforward, but it hides important differences.
Automation today includes:
These behaviors are not interchangeable. Treating them as one category makes the question appear larger and simpler than it really is. Without nuance, advertisers risk blocking legitimate activity, misallocating spend, or drawing the wrong conclusions about performance.
“Traffic quality isn’t solved at a single layer. Pre-bid controls matter, but without click-level defense, advertisers, especially performance buyers, are leaving themselves exposed. Not all invalid traffic is fraud, but without layered protection, sophisticated attacks can slip through.”
Utshob Alam, HUMAN Director of Research
Context Matters: What Internet-Scale Visibility Shows in Media Click Traffic
At HUMAN, we have been protecting against automation since our inception. AI-driven media traffic is quickly changing the programmatic landscape and it requires adaptable defenses. Sitting at the heart of digital media, HUMAN operates across a broad set of advertisers, publishers, and platforms. This allows us to see patterns emerge earlier and with more context than views limited to specific surfaces or moments in the media lifecycle.
Looking specifically at media click traffic across HUMAN’s Ad Click Defense customers, invalid traffic is clearly present but headline figures without context can be misleading. Rather than relying on a single aggregate percentage that implies universal impact, it is more accurate to say that invalid click rates vary widely by advertiser, channel, and traffic mix.
Historically, much of this automated activity has come from known crawlers and creative scanners – systems that interact with ads by design and are not inherently malicious. While recent narratives emphasize agentic AI, it undermines the incremental nature of the change.
In a joint whitepaper with LinkedIn, we discuss these nuances in depth and what is required to distinguish invalid traffic types effectively. In this case, by layering real-time behavioral analysis and click validation to native safeguards, LinkedIn improved invalid traffic detection over the first four months.
How HUMAN Distinguishes AI and Agentic Traffic Today
Across HUMAN’s Advertising Protection solutions, AI- and agent-driven interactions are classified under existing industry-standard invalid traffic categories.
The Human Defense Platform sees tens of millions of AI-associated bid requests a week, all of which are filtered as GIVT. Declared automation, including every major agentic provider as well as known crawlers, is flagged as general invalid traffic, while undeclared or deceptive automation is flagged as sophisticated invalid traffic.
This approach maximizes confidence that:
HUMAN’s ability to do this reliably is rooted in unmatched measurement coverage and end-to-end visibility—from ad delivery through onsite outcomes. That perspective is strengthened by AgenticTrust, part of HUMAN Sightline, which extends protection beyond media into account creation, authentication, and ecommerce flows.
AgenticTrust observes hundreds of thousands of AI agent-based requests every day, delivering on-site visibility. By understanding how agentic systems behave across the entire customer journey, not just where they touch ads, we gain insight that informs how agentic activity is detected and measured. This continuous feedback loop strengthens our defenses while keeping trust at the core.
Importantly, while agentic behavior is an emerging risk vector, it is not currently a primary driver of abuse compared to established fraud tactics. It is an area we actively monitor, test, and evolve against, without overstating its impact.
Media Signals and Onsite Outcomes Differ
The most important takeaway in this discussion to best understand AI-driven media impact is where that AI and agentic behavior appears.
In media environments, crawler and scanner activity is highly visible because those systems interact directly with ads. Onsite, agentic behavior looks different. AI tools may summarize content, compare products, or complete tasks without following a traditional user journey. In many cases, they do not generate meaningful engagement or downstream conversions.
Our recent analysis of holiday shopping traffic revealed how agentic systems generate distinct behavioral patterns across media environments and onsite activity. These agents interact with advertising differently than they navigate commerce experiences, reinforcing why media-side detection requires specialized visibility into ad delivery and engagement patterns, not just web traffic analysis.
Because of this, differences between media-reported activity and onsite outcomes are expected. They reflect the different roles agentic systems play across environments, not a contradiction in the data.
Why Precision Builds Trust
As AI capabilities accelerate, there is a temptation to rely on broad labels and large aggregate numbers. That approach may attract attention, but it does not help advertisers make better decisions.
Precision does.
Advertisers benefit when they understand:
This is how the industry moves beyond headlines and toward real understanding. It is also how trust is maintained as the ecosystem becomes more complex.
Moving Forward
AI and agentic systems will continue to shape how the web works and the digital media ecosystem that supports it. The goal is not to eliminate automation, but to measure it clearly and respond to it responsibly.
That requires accuracy, transparency, and evidence. It also requires being honest about what we know today and what is still evolving.
Cutting through the noise means resisting oversimplification and staying grounded in reality. That is how advertisers can move forward with confidence, not confusion.In a world saturated with AI-driven noise, clarity is the ultimate differentiator. Explainability is not just a safeguard. It is how trust scales as the ecosystem grows more complex. Connect with your HUMAN team today to better understand how agentic activity and AI might be affecting your media strategies. Together, we can turn opaque supply into transparent, high-performance channels where media dollars work harder.
Get visibility and control over AI agents and agentic browsers on your website.
