Invalid traffic — ad activity that doesn’t come from a real person with genuine interest — wastes ad budgets and erodes trust. We’re using large language models to keep advertisers, publishers and users even safer on our platforms.

Invalid traffic (IVT) is ad activity that doesn’t come from a real person with genuine interest. It wastes ad budgets, siphons revenue from publishers and erodes trust.

We’ve long used AI to fight IVT, and new applications are further improving our results. Our Ad Traffic Quality team, with Google Research and Google DeepMind, recently introduced industry-leading defenses powered by large language models to more precisely identify ad placements generating invalid behaviors.

Our new applications provide faster and stronger protections by analyzing app and web content, ad placements and user interactions. For example, they’ve significantly improved our content review capabilities, leading to a 40% reduction in IVT stemming from deceptive or disruptive ad serving practices ¹ . This helps advertisers better reach their intended audiences and keeps policy violators off our platforms.

Beyond this, we continue to run extensive automated and manual checks to ensure advertisers aren’t charged for IVT, even if an ad serves. These ad traffic quality innovations represent the latest in our two decades of work to defend against evolving threats and uphold the integrity of the digital ad ecosystem.

Source: Google Ads & Commerce Blog