Google now targets bad ads at bad actors

Google said Thursday it will block a record 8.3 billion ads globally in 2025 – up from 5.1 billion in the previous year. But the company has suspended far fewer advertiser accounts than that increase might suggest, raising questions about how its platform is managed.

The search giant attributed this disparity to its increasing use of artificial intelligence, particularly Gemini models – Google’s family of artificial intelligence systems – which Google says allow it to detect and block policy-violating ads earlier and more accurately. The company said that its artificial intelligence systems captured more than 99% of these ads last year before displaying them to users.

Both findings come from Google’s 2025 Ad Health Report and together reflect a broader change in implementation. While the most problematic ads are being stopped, fewer advertiser accounts are being suspended – marking a move from blocking bad actors outright to blocking individual ads on a case-by-case basis.

Google said the rise in blocked ads also reflects the increasing use of artificial intelligence by fraudsters to produce deceptive content at scale, with its Gemini models helping to spot patterns across large campaigns and block them early.

The shift also reflects a broader push by Google to integrate its Gemini models Deeper into its core products infrastructure, Including advertisingThe company is increasingly using artificial intelligence to automate campaign creation, detect policy violations, and respond to emerging threats in real time.

The company said that of the banned ads and suspended accounts, 602 million ads and 4 million advertiser accounts were linked to fraud.

Google removed more than 1.7 billion ads and suspended 3.3 million advertiser accounts in the United States in 2025, with ad network abuse, misinformation, and sexual content among the most common violations. In India, Google’s largest market by number of users, Google blocked 483.7 million ads – nearly double the previous year – even as account suspensions fell to 1.7 million from 2.9 million, with trademark, financial services and copyright issues among the top violations.

In a virtual press conference, Kirat Sharma, vice president and general manager of ad privacy and safety at Google, told reporters that the company has shifted toward more targeted, AI-driven execution “at a more granular level, at the creative level, rather than using a more explicit tool, like advertiser feedback.” He added that this approach helped reduce incorrect suspensions by 80% year-on-year.

Google’s multi-layered defenses, including advertiser verification (a process that requires companies to confirm their identity before running ads), are designed to prevent bad actors from creating accounts in the first place, Sharma said, adding that this contributed to the decline in suspensions.

Sharma said the numbers are likely to fluctuate over time as Google rolls out new defenses and bad actors adapt, as the company aims to stop malicious ads as soon as possible.