Google’s AI Crackdown on Ad Fraud: 39 Million Accounts Suspended in 2024
▼ Summary
– In 2024, Google suspended over 39 million advertiser accounts, tripling the previous year’s number, by enhancing its AI-driven ad fraud detection.
– Google implemented more than 50 updates to its AI safety systems, using large language models to detect fraud patterns that humans might miss.
– The majority of suspended accounts were blocked before any ads went live, showcasing a significant preemptive action against fraud.
– The U.S. had the highest enforcement with 1.8 billion ads removed, while India saw 2.9 million accounts suspended and nearly 250 million ads taken down.
– Despite AI’s role, over 100 Google staffers continue to work on tracing deepfake scams and refining AI responses, highlighting the ongoing need for human oversight.
Google’s war on ad fraud hit a new milestone in 2024, with the company suspending more than 39 million advertiser accounts, more than triple the number it took down in the previous year. The sharp escalation wasn’t due to a sudden surge in scammers, but rather a quiet shift in Google’s strategy: it leaned harder into AI. And the results are changing how fraud is hunted across the advertising world.
AI Moves to the Front Lines
For years, detecting ad fraud relied on a mix of reactive tools and manual reviews. But in 2024, Google rolled out more than 50 updates to its AI safety systems, putting large language models (LLMs) to work on spotting patterns that humans would miss. These models don’t just flag suspicious clicks or fake names—they analyze behavioral signals, compare payment details across campaigns, and identify business impersonation tactics before a single ad even runs.
The move paid off. According to Google’s internal data, a “vast majority” of the suspended accounts were blocked before they could go live with any ads. That level of preemptive action is rare in an industry where fraudsters often burn through accounts quickly and move on.
The U.S. saw the highest enforcement, with 1.8 billion ads removed and over 39 million advertiser accounts terminated. In India, 2.9 million accounts were suspended and nearly 250 million ads taken down. The top violations ranged from impersonating trusted brands to running unlicensed healthcare ads and misleading financial promotions.
Still a Human Problem, for Now
AI may be handling the bulk of detection, but Google hasn’t taken its eyes off the ball. The company says more than 100 staffers, drawn from Ads Safety, Trust and Safety, and DeepMind, worked on tracing back the source of deepfake scams that impersonated public figures. While the AI handles scale, human teams are still tasked with interpreting edge cases, dissecting new fraud trends, and refining how the systems respond.
Election-related ad enforcement was also in focus. With over 50 countries holding elections in 2024, Google verified nearly 9,000 new political advertisers and removed more than 10 million political ads that violated policies. Despite the high scrutiny, election ads made up a small slice of overall enforcement figures, suggesting the bulk of fraud remains financially motivated rather than politically driven.
The underlying message is clear: ad fraud isn’t going away, but AI has changed the balance of power. Google’s ability to detect and block fraud before it happens—at scale—is a leap forward, and it sends a message to the broader ad tech ecosystem: automated abuse requires automated defense.
Whether others will follow with the same level of investment remains to be seen. But for now, Google’s approach offers a glimpse into the future of ad safety, where smart algorithms don’t just clean up the mess, they prevent it from happening at all.
