Google Blocks 8.3 Billion Dangerous Ads by 2025 Thanks to Gemini's Help

Photo: Google

JAKARTA - Throughout 2025, Google managed to block or delete more than 8.3 billion ads and suspend 24.9 million advertiser accounts globally, thanks to Gemini-based tools.

In the Ads Safety Report 2025, Google claims that the integration of the AI model is able to detect more than 99% of malicious ads before they are shown to users.

The search giant said that the Gemini-based system is able to analyze hundreds of billions of signals, ranging from account age, campaign patterns, to behavioral indicators to identify threats more accurately.

From the total actions taken throughout 2025, Google also found 602 million ads and 4 million accounts related to fraudulent practices.

In Indonesia alone, Google noted that it had removed 191.9 million ads and suspended around 385 thousand advertiser accounts that violated the policy.

This step is claimed to be part of efforts to increase the security of the digital advertising ecosystem while maintaining user trust.

"While technology catches most policy violations, our extensive advertiser verification program provides another vital layer of prevention," said VP & General Manager, Ads Privacy and Safety, Keerat Sharma in his official blog.

Google also highlighted the increasing use of generative AI by criminals to create large-scale deceptive ads.

To compensate, Gemini is used to review most Responsive Search Ads instantly and block harmful content during the submission process, with plans to expand these capabilities to more ad formats.

"Gemini also allows us to process user input more efficiently, which helps our team take action on four times as many user reports in 2025 as the previous year," he added.

On the other hand, the increased accuracy of the model also reduces the suspension error for legitimate advertisers by up to 80%, so that credible businesses can still run their ad campaigns.

"By analyzing more than just image and text patterns, Gemini can better distinguish between credible offers and elaborate traps," he concluded.