Google's 2025 Ads Safety Report: Google blocked a record 8.3B ads, up 63% YoY, but suspended 36% fewer advertisers, attributing the disparity to its use of AI
Context & Ripple Effects
Google’s safety reporting shows ad enforcement operating at billion-ad scale for years: removals rose from 2.3 billion ads in 2018 to 3.4 billion in 2021 and 5.5 billion in 2023. Account suspensions had also climbed sharply, including 39.2 million U.S. advertiser accounts in 2024.
The 2025 report breaks that earlier coupling: blocked ads rose 63% to 8.3 billion while advertiser suspensions fell 36%. Google attributes the divergence to AI, making the story less about enforcement volume alone than about a shift toward detecting and stopping individual ad violations without necessarily removing the advertiser.
First-order effects
- Google is blocking substantially more policy-violating ads while suspending fewer advertiser accounts, altering the practical enforcement outcome for advertisers whose campaigns trigger safety systems.
- Google’s AI-based enforcement becomes more central to ads safety operations, with ad-level intervention apparently handling a larger share of violations.
Second-order effects
- Advertisers and agencies face greater need for policy-compliant creative and automated campaign controls, since individual ads can be blocked at higher volume even when an account remains active.
- Other major ad platforms will face pressure to demonstrate similarly scalable enforcement, while publishers and users may see more unsafe ads intercepted before distribution.
Third-order effects
- If this pattern persists, ad-safety enforcement may move from periodic account-level removals toward continuous, AI-assisted moderation of ads and advertisers at the level of individual submissions.
- That shift increases the importance of operational assurance: platforms will need to show that automated enforcement can scale without making policy decisions opaque or excessively error-prone.
The trend: This is one data point in the shift from account-centric trust-and-safety actions to AI-assisted, high-frequency enforcement across the advertising supply chain.