Google has published its 2025 Ads Safety Report, and the headline claim is significant: Gemini-powered systems helped stop more than 99% of policy-violating ads before they were shown to users. The company says these upgrades improved both speed and precision in identifying deceptive campaigns at scale.
According to the report, Google blocked or removed more than 8.3 billion ads and suspended 24.9 million advertiser accounts in 2025. It also cited election-related enforcement actions involving hundreds of millions of ads and millions of accounts tied to policy violations. Taken together, those numbers illustrate how AI has moved from a support role into the core of ad integrity operations.
Google describes Gemini as analyzing a broad mix of signals, including account behavior, campaign setup patterns, and indicators commonly associated with policy evasion. The objective is to identify abuse earlier in the lifecycle, before a campaign can gain reach. That shift from reactive moderation to proactive prevention is important as scammers use automation and synthetic content to iterate quickly.
For advertisers and publishers, stronger machine-led enforcement offers clear upside: less fraud, safer inventory, and potentially better long-term user trust in ad-supported products. But it also increases pressure on platforms to provide transparent policy guidance, predictable enforcement standards, and credible appeals for false positives. At internet scale, trust depends on both security outcomes and procedural fairness.
The strategic context is broader than one platform. As every major ad network applies more AI to trust and safety, competitive differentiation may come from who can combine strict enforcement with usable controls for legitimate businesses. In practical terms, the winners will likely be platforms that reduce harmful exposure without making compliant advertisers feel locked in a black-box process.
Why it matters
Digital advertising powers a large share of the web economy. If harmful ads are consistently blocked before delivery, that reduces consumer risk and fraud losses at global scale. The next challenge is proving that AI enforcement remains explainable, consistent, and accountable as systems become more autonomous.
Source: Google's 2025 Ads Safety Report post.