Google shifts focus to blocking harmful ads instead of just bad actors

Google said on Thursday that it blocked a record 8.3 billion ads worldwide in 2025, a sharp increase from the 5.1 billion ads it stopped the previous year. Despite that surge, the company suspended significantly fewer advertiser accounts than the scale of blocked ads might imply, highlighting a notable shift in how it enforces policies across its advertising ecosystem.

The company attributed this gap to its expanding use of artificial intelligence, particularly its Gemini family of models. According to Google, these systems enable earlier, more accurate detection of policy-violating content, allowing the company to stop problematic ads before they are ever displayed. Google said its AI-powered systems intercepted more than 99% of such ads before they reached users.

These insights come from the company's 2025 Ads Safety Report and reflect a broader evolution in enforcement strategy. While the volume of harmful or policy-violating ads being blocked has increased substantially, the number of advertiser account suspensions has declined. This suggests that Google is now focusing more on identifying and removing individual ads rather than taking broader action against entire advertiser accounts.

Google explained that the rise in blocked ads is also linked to scammers' growing use of generative AI tools, which now enable them to create misleading or fraudulent content at scale. The company said its Gemini models are being used to analyse patterns across large campaigns, making it easier to detect coordinated activity and stop harmful ads earlier in the process.

The change also aligns with Google's wider efforts to embed AI more deeply into its products and infrastructure. In advertising specifically, the company is increasingly relying on AI to automate campaign creation, identify policy violations, and respond to emerging threats in real time.

Among the enforcement actions taken in 2025, Google reported that 602 million ads and around 4 million advertiser accounts were connected to scams. In the United States, the company removed more than 1.7 billion ads and suspended 3.3 million advertiser accounts, with common violations including ad network abuse, misrepresentation, and inappropriate content.

In India, Google's largest market by user base, the company blocked 483.7 million ads — nearly double the previous year's total. At the same time, advertiser account suspensions in the country declined to 1.7 million from 2.9 million. The most frequent violations there involved trademark misuse, financial services issues, and copyright-related concerns.

Speaking during a virtual briefing, Keerat Sharma, vice president and general manager for ads privacy and safety at Google, said the company has moved toward more precise, AI-driven enforcement. He explained that the approach targets issues in detail — focusing on individual ad creatives — rather than relying on broader measures like account suspensions. According to Sharma, this shift has helped reduce incorrect suspensions by about 80% compared to the previous year.

Google also pointed to its layered defence systems, including advertiser verification requirements, which are designed to ensure that businesses confirm their identities before running ads. These measures aim to stop bad actors from entering the system in the first place, contributing to the overall decline in account suspensions.

Sharma noted that enforcement figures are likely to continue evolving as Google introduces new safeguards and as malicious actors adapt their tactics. The company's ongoing objective, he said, is to intercept harmful ads as early as possible in the advertising pipeline, preventing them from ever reaching users.