Google's AI blocked nearly half a billion harmful ads in India last year. Here's how

In a significant move to enhance online safety, Google has reported the removal of approximately 483.7 million harmful advertisements in India over the past year. This staggering figure highlights the scale of online misuse and the pivotal role that artificial intelligence (AI) plays in addressing it. According to the company's 2025 Ads Safety Report, Google not only blocked these ads but also suspended 1.7 million advertiser accounts for serious or repeated violations. The shift in Google's enforcement strategy is noteworthy, transitioning from a reactive moderation approach to a proactive detection model. This evolution is largely driven by advanced AI systems, including their latest model, Gemini. Google asserts that these models are designed to analyze user intent and behavioral signals, enabling early detection of harmful ads. "Bad actors are using generative AI to create deceptive ads at scale, and Gemini helps us detect and block them in real time," the company explained, revealing that over 99% of violating ads globally were intercepted at an early stage. On a larger scale, Google reported that it removed or blocked over 8.3 billion ads worldwide in 2025. In India, the majority of the violations were concentrated in key areas such as misleading financial services, trademark infringements, and the misuse of personalization tools for targeted manipulation. Notably, financial scams and impersonation tactics have emerged as major threats, with malicious actors continuously adapting their strategies to evade detection, often mimicking trusted brands or exploiting user data. In addition to removing specific ads, Google emphasized the importance of suspending repeat offenders. The company has taken decisive action by shutting down 1.7 million advertiser accounts in India that demonstrated a pattern of policy violations. Google's commitment to combating scams is reinforced by its use of advanced AI to identify and neutralize threats, with Gemini enhancing these capabilities further. "Our models analyze hundreds of billions of signals, including account age, behavioral cues, and campaign patterns, to stop threats before they reach the public," Google stated. The new models offer a deeper understanding of intent, allowing for the preemptive blocking of malicious content, even when designed to avoid detection.