Google has ramped up its utilization of Gemini AI models to combat harmful ads on its advertising platforms, as scammers and threat actors continuously evolve their tactics to evade detection.

In a recent update, Google revealed that it has successfully blocked or removed 8.3 billion ads and suspended 24.9 million advertiser accounts in 2025, with 602 million ads associated with scams being taken down.

Malvertising remains a significant issue on Google’s ad network, with attackers utilizing ads that mimic legitimate brands and services to distribute malware, steal cryptocurrency, or direct users to phishing sites.

These malicious campaigns often employ cloaking techniques and URL redirects to appear as trusted sites, including leveraging Google’s domains and those of legitimate software downloads and authentication portals.

Recent malicious campaigns highlighted by BleepingComputer include fake login pages targeting Google Ads accounts, distributing malware through ads posing as tools like Google Authenticator, and displaying ads for fraudulent cryptocurrency platforms that drain visitors’ wallets.

Google reports that cybercriminals are now leveraging generative AI in their campaigns, enabling them to rapidly scale and execute more sophisticated operations.

“Bad actors are leveraging generative AI to produce deceptive ads at scale, and Gemini aids us in real-time detection and blocking. By the end of the previous year, the majority of Responsive Search Ads in Google Ads underwent instant review, with harmful content being blocked upon submission – a feature set to expand to more ad formats this year,” explained Keerat Sharma, VP & General Manager, Ads Privacy and Safety.

To counter these threats, Google emphasizes its reliance on Gemini AI-powered systems to automate the identification and prevention of malicious ads before they reach users.

While previous detection systems focused on analyzing keywords for malicious intent, Gemini can now analyze billions of signals, including advertiser behavior, account history, campaign patterns, and intent, to assess the malicious nature of an ad.

In the US, Google removed 1.7 billion ads and suspended 3.3 million advertiser accounts in 2025, with “abusing the ad network” and “misrepresentation” ranking as the top policy violations.

Google’s integration of artificial intelligence has significantly enhanced its ability to address malicious and scam ads that slip through initial reviews, enabling faster processing of user reports compared to previous years.

Google also notes that the increased accuracy of its AI models has led to an 80% reduction in incorrect advertiser suspensions.

The company plans to further expand Gemini’s implementation across various ad formats and enforcement systems, with the goal of preemptively blocking malicious campaigns during the submission phase.