Google blocked over 8 billion ads with AI assistance in 2025

Google presents the 2025 figures for blocked advertising and fraudulent user accounts: Google blocked over 8.3 billion ads and around 25 million advertiser accounts last year.

In a blog post, Google provides a brief overview. According to the post, employees have been using advanced AI for a long time to detect and slow down fraudsters. Gemini takes this a step further. Google uses it to examine hundreds of billions of signals, including account age, behavioral characteristics, and campaign patterns, to block threats before they reach people. Unlike older systems that relied on keywords, the new models better understand intent, allowing them to detect and proactively block malicious content – “even when it’s designed to evade detection.” According to its own data, Google now blocks over 99 percent of policy-violating ads before they are even served.

In 2025, Google blocked or removed over 8.3 billion ads, restricted over 4.8 billion ads, closed more than 24.9 million ad accounts, and blocked or restricted 480 million websites. The brief overview also shows a breakdown of blocked ads by “offenses” against specific policies, with the largest category being 1.29 billion ads that misuse the advertising network. Of the 480 million blocked websites, AI-powered control systems detected 97 percent (over 467 million).

Google also provides figures for the EU

A separate report analyzes last year's figures for the European Union. According to this, Google removed 1.6 billion ads in the EU and blocked 2 million ad accounts. Here too, the misuse of the advertising network is the number one policy violation.

Google aims to be able to react in real-time in the race against malicious actors. Perpetrators use generative AI to create fraudulent ads on a large scale, and Gemini AI helps to detect and block them in real-time. Last year, the majority of so-called Responsive Search Ads created in Google Ads were analyzed immediately, and dangerous content was blocked before transmission. This is planned to be extended to other ad formats this year.

Gemini has also helped to manage user feedback more efficiently. Teams were able to take four times as many countermeasures against reported ads as in the previous year. If a threat does slip through, Google can now neutralize it faster. Another success was achieved with false positives. Google was capable of reducing the number of incorrectly blocked advertisers by 80 percent in 2025.

Google's Ads Safety Reports are published regularly. Google has been blocking billions of ads every year since the early 2020s according to reports. Competitor Meta has a reputation for being significantly more hesitant in suppressing fraudulent advertising. This seems to pay off for the company.

(dmk)