Google leveraged AI to suspend ad accounts—a staggering 39.2 million of them—in 2024, marking the tech giant’s most aggressive crackdown on ad fraud to date.

This number more than triples the previous year’s suspensions and highlights Google’s growing reliance on artificial intelligence to protect its advertising ecosystem. The company said it used large language models (LLMs) to detect red flags like business impersonation and fake payment credentials, helping it preemptively disable fraudulent ad accounts before they ever went live.

“AI models have driven dramatic safety gains,” said Alex Rodriguez, General Manager for Ads Safety at Google, during a media briefing. “But human oversight remains critical.”

Rodriguez revealed that more than 100 experts—including teams from Google’s Ads Safety, Trust and Safety, and DeepMind divisions—worked together to tackle deepfake ad scams, particularly those involving impersonated public figures. Their collaboration led to technical safeguards and 30+ policy updates in 2024 alone.

Thanks to these advancements, over 700,000 high-risk advertising accounts were suspended, resulting in a 90% drop in deepfake ad reports, according to Google.

The numbers are even more staggering at a country level. In the United States, Google blocked 1.8 billion ads and suspended 39.2 million advertiser accounts, mainly for ad network abuse, healthcare misinformation, trademark violations, and deceptive ad content. In India, the second-largest internet market after China, 2.9 million accounts were banned and over 247 million ads taken down—mostly linked to financial scams, gambling, and personalized ad violations.

In total, Google says it blocked 5.1 billion ads, removed 1.3 billion pages, and restricted 9.1 billion more in 2024. That’s down from 5.5 billion ads and 2.1 billion publisher page actions in 2023, signaling improvements in preemptive detection.

The company also confirmed the verification of 8,900 new election advertisers globally and the removal of 10.7 million election-related ads. Still, Rodriguez emphasized that political ads are just a fraction of Google’s overall ad volume and do not significantly affect its safety metrics.

While the scale of automated enforcement is unprecedented, concerns about fairness persist. Google assured advertisers that every suspension can be appealed, and that a human review team ensures each case is treated with transparency and precision.

“Often, the issue was a lack of clarity in our messaging,” Rodriguez admitted. “We’ve since upgraded our transparency tools to give advertisers more detailed explanations and clearer appeal pathways.”

With AI continuing to reshape the digital advertising landscape, Google’s latest moves underscore the balance between automation and accountability in the fight against fraud.

Get the Latest AI News on AI Content Minds Blog