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
Google said Thursday it blocked a record 8.3 billion ads globally in 2025 — up from 5.1 billion the year before.
Context & Ripple Effects
Google’s ad-safety disclosures have long tracked both ad removals and account enforcement, with prior coverage showing multi-billion annual removals and a sharp increase in advertiser suspensions in 2024. The 2025 report breaks that recent pattern: blocked ads rose sharply while advertiser suspensions fell.
That divergence matters because Google explicitly links it to AI, suggesting enforcement is being applied more granularly at the ad level rather than relying as heavily on account-level exclusion.
First-order effects
- Google removes or blocks substantially more ads while suspending fewer advertisers, changing the immediate balance of enforcement toward individual creative or campaign decisions.
- Advertisers whose ads trigger policy systems may face more ad-level interventions without necessarily losing their entire accounts; Google must operate and review a much larger volume of such decisions.
Second-order effects
- The shift raises the importance of advertisers’ policy-compliance workflows, since more campaigns can be interrupted or rejected even when the advertiser remains active on the platform.
- Other large ad platforms face pressure to demonstrate that automated enforcement can scale against harmful or noncompliant ads without indiscriminately shutting out legitimate advertisers.
Third-order effects
- If this pattern persists, AI-based moderation could move ad safety from episodic account takedowns toward continuous, high-volume controls over ads and campaigns.
- The trade-off will increasingly center on the accuracy, consistency, and contestability of automated enforcement, especially as platforms depend on it for decisions affecting access to advertising markets.
The trend: This is one data point in the expansion of AI as the operational layer for platform policy enforcement at internet-scale volumes.