SAN FRANCISCO, United States, April 16, 2026 (BSS/AFP) - Google on Thursday said tools utilizing Gemini artficial intelligence helped it derail some 8.2 billion online ads which violated its policies last year.

"We really dove into how to rethink our systems using AI," Google's vice president of ads privacy and safety Keerat Sharma said during a briefing with journalists.

"Our latest models better understand intent, helping us spot malicious content and preemptively block it, even when it's designed to evade detection."

An annual Ads Safety Report released by Google indicated that the internet giant's systems caught more than 99 percent of ads that broke its rules before they were shown to people online.

"Our goal is to stop badness before it is ever exposed to any user," Sharma said.

Google made 35 changes to its ad policies last year in updates intended to help keep pace with evolving trends in scam marketing, according to Sharma.

Gemini-enhanced ad system tools analyze hundreds of billions of "signals," such as the age of advertiser accounts, how they are acting, and patterns in campaigns to determine if something is amiss, he explained.

"We took a step back and really re-thought how we wanted to build ad safety from a ground-up perspective," Sharma said.

"With Gemini, we've been able to take a step back and combine hundreds of billions of signals to get a much better understanding of what an advertiser's intent is based on the combination of everything that we know about the advertiser," Sharma said.

"All of that put together gives us this really nuanced perspective as to whether the intent of a specific ad is positive or potentially harmful."

While bad actors are using generative AI to create deceptive ads at scale, Gemini helps detect and block them in real time, according to Google.

Ad tactics barred by Google policies include deepfakes of public figures or misappropriation of their genuine images.

Another layer to Google's ad defense is verifying advertisers, which account for more than 90 percent of ads served on Google, according to Sharma.

"Verification is a tool that we use to ensure that an advertiser is indeed a legitimate entity," Sharma said.

"We're continuing to evolve our defenses to stay ahead of even the most advanced schemes." he said.