Google has revealed that its AI systems, powered by Gemini, successfully blocked over 99% of policy-violating ads before they were shown to users in 2025. The update highlights Google’s continued push to strengthen its ad safety ecosystem as scam tactics become increasingly sophisticated.
Gemini Enhances Ad Safety with Intent-Based Detection
According to Google, Gemini-powered systems significantly improve how malicious ads are identified and stopped. Key improvements include:
- Intent-based detection: Goes beyond keywords to understand advertiser behavior
- Real-time analysis: Processes hundreds of billions of signals
- Early blocking: Stops harmful ads before they enter the delivery pipeline
Signals analyzed include account age, campaign patterns, and behavioral indicators. This allows the system to detect scams even when they attempt to bypass traditional filters.
Massive Enforcement Actions in 2025
Google’s latest Ad Safety Report outlines the scale of enforcement powered by Gemini AI:
- 99%+ of violating ads blocked before serving
- 8.3 billion ads blocked or removed
- 24.9 million advertiser accounts suspended
- 602 million scam-related ads removed
- 4 million scam-linked accounts banned
Faster Review Systems & Improved Accuracy
Google also introduced improvements to review speed and accuracy:
- Instant review for most Responsive Search Ads
- Expansion planned for more ad formats
- 4× increase in user report processing
- 80% reduction in incorrect advertiser suspensions
Additionally, advertiser verification systems now add identity checks before ads go live, helping prevent bad actors from entering the ecosystem.
Tackling AI-Generated Scams
With the rise of generative AI, Google says Gemini is now capable of:
- Detecting AI-generated deceptive ads
- Identifying advanced scam patterns in real time
- Strengthening automated enforcement across ad formats
Google plans to further expand intent-based detection models, real-time enforcement systems, and advertiser verification processes. The company aims to balance strong fraud prevention with continued support for legitimate advertisers.
