The AI Advertising Measurement Problem Is Really About Governance

The advertising industry has spent recent years rebuilding its measurement systems around a shrinking pool of identifiers. Now another complication has emerged: artificial intelligence is changing not just how ads are delivered, but who — or what — sees them before a purchase ever happens.

That's why the Interactive Advertising Bureau (IAB) is building a common framework for measuring ads shown to, or influencing, AI agents, with initial findings expected in November. The project aims to bring clarity to questions traditional attribution — the method of tying ad exposure to a purchase — was never designed to answer. For instance, if an AI agent researches products, encounters commercial messaging, and ultimately helps a user complete a purchase, the industry will need to determine what counts as an ad impression in that scenario, which signals confirm influence, and how credit gets divided across the ecosystem.

These questions might sound theoretical, but the commercial stakes are rising fast as AI platforms stop being mere consumer interfaces and start building advertising businesses of their own. OpenAI illustrates this clearly: just six months after launching ads in the US, the company rolled out ChatGPT Ads to several international markets and then to 31 more countries across Europe. Along the way, it built shopping tools, introduced a cost-per-click pricing model, hired seasoned advertising professionals, and forged relationships with agencies and ad tech partners.

Establishing standards, though, is only part of the challenge. The more important question is who controls the underlying signals. If AI platforms retain the data needed to prove how their systems influenced a transaction, independent measurement could end up depending on information supplied by the very companies selling that advertising. Industry frameworks can set definitions and expectations, but their effectiveness will hinge on whether platforms grant enough access for those standards to be independently verified.

This situation already feels familiar. A similar debate has played out around the Attribution API proposed by the World Wide Web Consortium: proponents argue browser-based attribution is a necessary way to preserve measurement while curbing cross-site tracking, while critics worry that shifting this responsibility to browsers could concentrate influence in the hands of the companies that control them. The circumstances differ, but the governance question stays the same: large technology companies typically have far more technical resources, data, and staff to devote to lengthy standardization processes than smaller publishers, independent ad tech firms, or advertisers do.

Even so, almost no one disputes the industry's need for unified standards: shared requirements could reduce fragmentation and give advertisers and publishers clearer rules for operating in a fast-changing landscape. OpenAI's rapid commercialization shows that the AI advertising market will need such rules sooner than expected. But setting measurement standards will only be the first test. The next one is whether the resulting systems can deliver genuine transparency and interoperability — rather than simply transferring power from one generation of dominant tech platforms to the next.