Google Strengthens Ad Safety, Removes 3564 Violating URLs in Taiwan
Google released its "2025 Ad Safety Report," detailing how its AI model Gemini has enhanced fraud prevention and ad review. Globally, over 8.3 billion ads were removed and 24.9 million accounts suspended. In Taiwan, 3564 violating URLs were removed in compliance with local regulations.
📋 Article Processing Timeline
- 📰 Published: April 17, 2026 at 13:28
- 🔍 Collected: April 17, 2026 at 14:01 (32 min after Published)
- 🤖 AI Analyzed: April 17, 2026 at 17:36 (3h 34m after Collected)
Central News Agency
(Central News Agency reporter Wu Chia-hao, Taipei, 17th) Google today released its "2025 Ad Safety Report," strengthening its anti-fraud and ad review mechanisms through the AI model Gemini. Last year, it intercepted over 99% of violating content before ad delivery, cumulatively removing over 8.3 billion ads and suspending more than 24.9 million accounts. Google also complied with regulations in Taiwan, removing 3564 violating URLs.
Keerat Sharma, Google's Vice President and General Manager of Ads Privacy and Safety, stated in an official blog post that Google's team has long used advanced AI technology to identify and stop fraudsters, and Gemini takes this work to a higher level. Google's models analyze hundreds of billions of signals, including account usage time, behavioral cues, and ad campaign patterns, to intercept threats before they reach users.
Sharma emphasized that unlike earlier keyword-based systems, Google's latest models can more deeply understand intent, accurately identifying and preemptively blocking malicious content even when it is deliberately designed to evade detection.
In Taiwan, Google continues to comply with the "Fraud Crime Prevention Act" and takes strict measures to ensure platform safety. From July 2024 to June 2025, Google received 236 removal requests from the government in Taiwan and successfully removed 3564 URLs, including 227 URLs violating laws and 3337 URLs violating Google's anti-fraud policies.
Sharma stated that by analyzing information beyond image and text patterns, Gemini can more clearly distinguish between reliable offers and complex traps. This precise grasp of nuances helped Google reduce erroneous account suspensions by 80% last year. This accuracy allows Google to prioritize the removal of harmful content while helping compliant businesses successfully launch ads. (Editor: Lin Chia-hsien) 1150417
Choose to stand with facts, every sponsorship you make is a force to protect press freedom.
Download the Central News Agency "First-hand News" APP to stay updated with the latest news.
The text, images, and audio-visual content of this website may not be reproduced, publicly broadcast, or publicly transmitted and used without authorization.
(Central News Agency reporter Wu Chia-hao, Taipei, 17th) Google today released its "2025 Ad Safety Report," strengthening its anti-fraud and ad review mechanisms through the AI model Gemini. Last year, it intercepted over 99% of violating content before ad delivery, cumulatively removing over 8.3 billion ads and suspending more than 24.9 million accounts. Google also complied with regulations in Taiwan, removing 3564 violating URLs.
Keerat Sharma, Google's Vice President and General Manager of Ads Privacy and Safety, stated in an official blog post that Google's team has long used advanced AI technology to identify and stop fraudsters, and Gemini takes this work to a higher level. Google's models analyze hundreds of billions of signals, including account usage time, behavioral cues, and ad campaign patterns, to intercept threats before they reach users.
Sharma emphasized that unlike earlier keyword-based systems, Google's latest models can more deeply understand intent, accurately identifying and preemptively blocking malicious content even when it is deliberately designed to evade detection.
In Taiwan, Google continues to comply with the "Fraud Crime Prevention Act" and takes strict measures to ensure platform safety. From July 2024 to June 2025, Google received 236 removal requests from the government in Taiwan and successfully removed 3564 URLs, including 227 URLs violating laws and 3337 URLs violating Google's anti-fraud policies.
Sharma stated that by analyzing information beyond image and text patterns, Gemini can more clearly distinguish between reliable offers and complex traps. This precise grasp of nuances helped Google reduce erroneous account suspensions by 80% last year. This accuracy allows Google to prioritize the removal of harmful content while helping compliant businesses successfully launch ads. (Editor: Lin Chia-hsien) 1150417
Choose to stand with facts, every sponsorship you make is a force to protect press freedom.
Download the Central News Agency "First-hand News" APP to stay updated with the latest news.
The text, images, and audio-visual content of this website may not be reproduced, publicly broadcast, or publicly transmitted and used without authorization.
FAQ
What AI model does Google use to enhance ad safety?
Google uses the AI model Gemini to strengthen its anti-fraud and ad review mechanisms.
How many violating URLs did Google remove in Taiwan last year?
Google removed 3564 violating URLs in Taiwan last year.