Google's Gemini Blocks Ads

Google said Gemini-powered systems blocked or removed more than 8.3 billion ads in 2025, with more than 99% stopped before users saw them. (blog.google) The company said it suspended 24.9 million advertiser accounts last year and flagged 602 million scam-linked ads and 4 million scam-linked accounts. Its 2025 Ads Safety Report says Gemini now analyzes signals such as account age, behavior patterns and campaign activity, rather than relying mainly on keyword matching. (blog.google) (services.google.com) Google said Gemini also helped it process four times as many user reports as the year before and cut incorrect advertiser suspensions by 80% in 2025. By the end of the year, Google said most Responsive Search Ads were being reviewed instantly at submission. (services.google.com) (blog.google) Online ad policing has shifted from catching bad ads after they appear to screening them before they run. Google said newer scam campaigns are increasingly built with generative artificial intelligence, which lets fraudsters produce deceptive ads at scale. (blog.google) That shift also changed how Google describes enforcement. In 2024, the company said it blocked or removed 5.1 billion ads; in 2025, that figure rose to 8.3 billion, while Google said its systems were getting better at identifying intent and focusing action on bad actors earlier in the process. (blog.google) (services.google.com) The 2025 report breaks out where Google says it took the most ad action. The largest categories included 1.29 billion ads for abusing the ad network, 755 million for personalization violations and 646.7 million for legal requirements, with scam defenses accounting for 602 million removed ads. (services.google.com) Google said it also took action beyond ads themselves, restricting 4.8 billion ads, blocking or restricting more than 480 million web pages and acting against more than 245,000 publisher sites in 2025. The company said artificial intelligence-driven systems contributed to enforcement on more than 467 million of those pages. (services.google.com) In South Korea, Google said it removed 175.5 million policy-violating ads in 2025 and suspended 326,000 advertiser accounts. The Korea Herald, citing Google, reported copyright infringement was the biggest violation category there, followed by abuse of the ad network, dating and companionship services, misrepresentation and sexually explicit content. (koreaherald.com) Google said advertiser verification remains a second layer of defense alongside Gemini, aimed at confirming who is buying ads before campaigns go live. The company’s latest report frames 2025 as a year when more of that enforcement happened before an ad ever reached a screen. (services.google.com)