Google blocks record 8.3 billion ads in 2025 as AI targets bad ads over bad actors

At a glance:

• Google blocked 8.3 billion ads globally in 2025, up from 5.1 billion in 2024.
• AI-driven detection with Gemini models stopped 99% of policy-violating ads before users saw them.
• Advertiser account suspensions fell sharply, signaling a shift to granular, ad-level enforcement.

AI-powered enforcement reshapes Google's ad safety strategy

Google's 2025 Ads Safety Report reveals a dramatic surge in blocked advertisements — 8.3 billion globally, a 63% increase over the previous year. Yet the number of suspended advertiser accounts did not rise proportionally, pointing to a fundamental shift in how the company polices its platform. Instead of broadly suspending entire accounts, Google is now leveraging its Gemini AI models to detect and block individual policy-violating ads with greater precision.

This AI-driven approach allowed Google to stop more than 99% of problematic ads before they reached users, the company said. The strategy reflects both the growing sophistication of generative AI tools used by scammers to mass-produce deceptive content and Google's own deepening integration of AI across its core infrastructure. By identifying patterns across large campaigns, Gemini can intervene earlier in the ad lifecycle, reducing the need for blunt-force account suspensions.

Fewer suspensions, more targeted action

The disparity between blocked ads and suspended accounts is striking. While ad blocks jumped by over 3 billion, the number of advertiser accounts suspended globally fell to 4 million from 5.1 million the year before. In the U.S., Google removed 1.7 billion ads and suspended 3.3 million accounts; in India, its largest user market, it blocked 483.7 million ads — nearly double the prior year — while suspensions dropped to 1.7 million from 2.9 million.

Keerat Sharma, VP and general manager of ads privacy and safety at Google, explained that the company is now focusing on "granular, creative-level" enforcement rather than sweeping account bans. This shift has cut incorrect suspensions by 80% year over year. Sharma emphasized that layered defenses, including advertiser verification processes that require businesses to confirm their identity, are designed to prevent bad actors from creating accounts in the first place, further reducing the need for suspensions.

Scam ads and regional enforcement patterns

Among the blocked ads and suspended accounts, 602 million ads and 4 million advertiser accounts were linked to scams. The most common violations varied by region: in the U.S., ad network abuse, misrepresentation, and sexual content topped the list; in India, trademark, financial services, and copyright issues were most prevalent. The rise in blocked ads also reflects the growing use of generative AI by scammers to produce deceptive content at scale, with Google's Gemini models helping detect and block such patterns earlier.

Sharma noted that enforcement numbers are likely to fluctuate as Google rolls out new defenses and bad actors adapt. The company's goal remains to stop harmful ads as early in the pipeline as possible, using AI not just to react to violations but to anticipate and prevent them. This evolving strategy underscores the central role AI now plays in maintaining the integrity of digital advertising at global scale.

Looking ahead: AI as both shield and sword

As generative AI tools become more accessible, the arms race between platforms and bad actors intensifies. Google's pivot to AI-driven, ad-level enforcement represents a significant evolution in digital safety — one that prioritizes precision over punishment. By catching violations earlier and more accurately, the company aims to protect users without unduly penalizing legitimate advertisers.

However, the effectiveness of this approach will depend on Google's ability to stay ahead of increasingly sophisticated scams and the rapid evolution of AI-generated content. The 2025 report suggests that while the volume of bad ads is rising, so too is the capability to stop them — a dynamic that will likely define the future of online advertising safety.