On a global scale, Google removed over 8.3 billion harmful ads and suspended 24.9 million advertiser accounts last year, as reported in Google’s ‘2025 Ads Safety Report’ released on Thursday.
The company noted that the integration of its Gemini AI models has significantly enhanced its ability to identify and eliminate malicious actors in real-time, particularly as scammers increasingly leverage generative AI to produce deceptive advertisements at scale.
Out of the 8.3 billion ads removed worldwide, Google reported that over 99% were stopped before users ever encountered them.
”Our safety teams operate continuously to combat bad actors employing increasingly advanced, harmful ads,” stated Keerat Sharma, VP & General Manager, Ads Privacy and Safety at Google, in a blog post.
”Our models analyze hundreds of billions of signals — such as account age, behavior patterns, and campaign trends — to thwart threats prior to reaching users. Unlike past keyword-based systems, our latest models have a better understanding of intent, allowing us to detect malicious content and proactively block it, even if it’s crafted to avoid detection,” Sharma added.
In India, the leading five policy violations responsible for ad removals were tied to trademarks, financial services, copyright infringements, personalization violations, and abuse of the ad network.
According to Google, Gemini vastly improved the processing of user reports, enabling action on four times as many reports in 2025 compared to the previous year.
