Google AI Blocked 8.3 Billion Bad Ads in 2025

Google blocked 8.3 billion fraudulent ads in 2025 using its Gemini AI model, with 99% of violations intercepted before reaching users. In the EU alone, 1.6 billion ads were removed and 2 million advertiser accounts suspended, mostly for network abuse and trademark violations. As bad actors increasingly use generative AI to evade detection, Google deployed Gemini to analyse hundreds of billions of signals, directly linking 602 million removed ads to scams. The system also reduced wrongful suspensions of legitimate advertisers by 80% and handled four times more user complaints than the previous year.