Agentic AI to Expose Flaws in Ad Measurement, Warns IAB

What happened

The Interactive Advertising Bureau (IAB) anticipates that the rise of agentic AI will highlight existing deficiencies in advertising measurement and attribution. Shailin Gabilan, the IAB's Vice President of Emerging Technology, noted that while these advanced AI systems offer significant opportunities for efficiency and personalisation in advertising, they will simultaneously expose long-standing weaknesses in how advertising effectiveness is tracked and evaluated. This development suggests a critical need for the advertising industry to re-evaluate and enhance its measurement frameworks to adapt to a landscape increasingly shaped by autonomous AI agents.

Why it matters

For business leaders and AI practitioners, this forecast from the IAB underscores a crucial governance challenge. The deployment of agentic AI, capable of making autonomous decisions and executing tasks, introduces new complexities in accountability and transparency, particularly in areas like advertising where impact and return on investment are paramount. If current measurement methodologies are already insufficient, the opaque nature of some AI operations could further obscure the true influence of campaigns, complicating ethical attribution and potentially leading to misallocation of resources. Addressing these "old cracks" now, before agentic AI becomes fully pervasive, is essential for maintaining trust, ensuring fair competition, and enabling responsible AI adoption across the marketing and advertising ecosystem.

The Institute take

The IAB's observation about agentic AI revealing existing flaws in ad measurement is more than just an industry-specific technical challenge; it’s a bellwether for AI governance across all sectors.

This scenario highlights the pervasive issue of "garbage in, garbage out" in a new light: if our foundational data and measurement systems are already compromised, deploying sophisticated AI agents on top of them will only amplify existing biases and inefficiencies, not solve them. Responsible AI leaders must recognise that advanced AI isn't a magic bullet for legacy problems; instead, it acts as a stress test for existing infrastructure and governance. The imperative is to first audit and fortify the underlying data and measurement frameworks before scaling AI agent deployment, ensuring that the automation doesn't merely accelerate poor practices but instead builds upon a robust, ethical foundation.