Google adds ad-safety AI

Google is adding new Gemini-powered safety tools to Ads Advisor as it says its systems blocked or removed more than 8.3 billion ads in 2025. (blog.google) Google said on April 16 that Gemini-powered tools stopped more than 99% of policy-violating ads before they ran, and suspended 24.9 million advertiser accounts. The company said 602 million blocked ads and 4 million suspended accounts were tied to scams. (blog.google) The new product update adds three “agentic” safety and policy features inside Ads Advisor, Google’s Gemini-based assistant in Google Ads. Google says the tools are meant to protect accounts and streamline approvals for advertisers working inside the ad platform. (blog.google) Ads Advisor is a chat-style tool in Google Ads that can answer account questions, troubleshoot ad serving, explain policy violations and apply suggested changes with the advertiser’s approval. Google began rolling it out to English-language advertisers in late 2025. (support.google.com, blog.google) That puts compliance work closer to campaign creation. In Google Ads, most ad reviews are completed within one business day, but edits can restart review and more complex checks can take longer. (support.google.com, support.google.com) Google’s pitch is that the same AI system can now do two jobs at once: catch scams before users see them and help legitimate advertisers fix policy problems faster. The company said Gemini looks at signals including account age, behavioral cues and campaign patterns to spot malicious intent. (blog.google) The scale of enforcement has risen from the prior report. Google said it blocked or removed 5.1 billion ads and suspended 39.2 million advertiser accounts in 2024, compared with 8.3 billion ads and 24.9 million accounts in 2025. (services.google.com, blog.google) Google has also said Gemini helps reduce mistaken suspensions, a long-running complaint from advertisers who can lose traffic while accounts or ads sit under review. Ads Advisor’s help page says users still need to verify that any suggested fix is accurate before applying it. (blog.google, support.google.com) The update leaves Google arguing that ad safety is no longer just a moderation layer after launch. It is becoming part of the product flow that decides whether an ad gets built, fixed and approved in time to run. (blog.google, support.google.com)