Gemini is stopping harmful ads before people ever see them

In 2025, Google's safety teams leveraged Gemini-powered tools to significantly enhance ad protection against increasingly sophisticated bad actors.
These tools analyzed hundreds of billions of signals—including account age, behavior, and campaign patterns—to detect and stop threats preemptively, catching over 99% of policy-violating ads before they served.
As a result, Google blocked or removed over 8.3 billion ads and suspended 24.9 million accounts, with 602 million ads and 4 million accounts linked to scams.
Gemini's advanced understanding of intent, beyond keyword matching, helped identify and block malicious content even when designed to evade detection.
The system also processed user feedback more efficiently, enabling action on over four times as many user reports year-over-year.
Additionally, Gemini's nuanced analysis reduced incorrect advertiser suspensions by 80%, allowing honest businesses to continue running ads while focusing on removing harmful content.
By the end of the year, most Responsive Search Ads were reviewed instantly, with harmful content blocked at submission—a capability being expanded to other formats.
This dual focus on precision and speed supported both user protection and legitimate business operations.