Google’s ad‑safety scale

Google says it blocked or removed more than 8.3 billion ads in 2025 and suspended 24.9 million advertiser accounts. (blog.google) The company said more than 99% of policy-violating ads were stopped before any user saw them, with Gemini-based systems helping reviewers flag scams and impersonation earlier in the process. (blog.google) (support.google.com) Google’s report put scams at 602 million blocked or removed ads and 4 million suspended accounts in 2025. In India, Google said it blocked or removed 483.7 million ads and suspended 1.7 million advertiser accounts; in South Korea, it said 175.5 million ads were removed. (blog.google 1) (blog.google 2) (koreaherald.com) The scale matters because Google’s ad business runs across Search, YouTube and millions of partner sites, and the company is trying to screen a flood of cheap, AI-made pitches before they enter that system. Google said large language models were used to spot patterns such as fake business identities and suspicious payment details. (blog.google) (support.google.com) Google also tied the report to a broader push on advertiser identity. In April 2025, it updated its Ads Transparency policy so users could see more information about the entity paying for an ad, not just the advertiser name. (support.google.com) Political ads are getting tighter scrutiny too. Google updated its political content rules in July 2025 and added European Union restrictions in September 2025, while continuing to require verification for many election advertisers and to publish records in its transparency tools. (support.google.com 1) (support.google.com 2) (support.google.com 3) The company said the newer systems also reduced mistaken account suspensions. Google said its 2025 enforcement work cut false suspensions by 80%, a key point for legitimate advertisers that can lose sales when campaigns are shut off by error. (blog.google) (ppc.land) Google publishes one of these safety tallies each year, but the 2025 numbers show the company shifting from chasing bad ads one by one to screening the people and payment trails behind them. The next test is whether those systems keep blocking fraud at this scale without sweeping in more legitimate advertisers. (blog.google) (support.google.com)