Google said it blocked 99 percent of harmful ads before they reached users, as it expands the use of its AI model Gemini to detect fraud and enforce advertising policies.
In its 2025 Ads Safety Report, the company said it removed or blocked 8.3 billion ads, with the vast majority stopped before being seen.
Google said artificial intelligence, combined with human review, played a central role in identifying and preventing policy violations across its platforms.
Among the removed ads, a large portion was linked to network abuse, while others involved issues such as personalization violations, legal non-compliance and misleading content.
The company also took action against advertisers and publishers, suspending nearly 24.9 million advertiser accounts and restricting hundreds of millions of web pages.
Google said its systems are designed to detect fraudulent ads before they appear, helping prevent scams such as fake investment offers and misleading promotions.
Keerat Sharma said generative AI is increasingly being used by bad actors to create deceptive content, but Gemini helps identify and block such threats in real time.
He added that the system analyses large volumes of data signals, including behaviour patterns and account activity, to stop malicious campaigns early.
Google said it updated its advertising policies multiple times during the year and plans to further expand Gemini-based protections across more ad formats in 2026.
Despite the increased use of AI, the company said human oversight remains essential in ensuring accurate and fair enforcement.
-20260416124550.webp)
-20260408113340.webp)
-20260403095538.webp)
-20260329132714.webp)
-20260422055630.jpg)
-20260416052446.webp)
