Google is leveraging AI to combat advertising fraud, addressing a major concern for banks and lawmakers regarding scam accounts. In 2024, the tech giant reported the suspension of 39.2 million advertiser accounts in the U.S.—more than three times the figure from the previous year. In total, nearly half a billion ads associated with scams were removed.
Last year, Google introduced over 50 updates to its Large Language Models, aimed specifically at identifying and eliminating bad actors. According to the company, these enhancements accelerated complex investigations, enabling them to detect fraud signals, such as false payment information, during account creation. This proactive approach has prevented billions of policy-violating ads from being displayed to consumers while allowing legitimate businesses to reach customers more efficiently.
Additionally, Google has adjusted its defenses against emerging scams that exploit deep fake AI technology to impersonate public figures. According to the Federal Trade Commission, impersonation scams targeting businesses and government officials cost Americans approximately $2.95 billion last year.
In response to this challenge, Google formed a team of over 100 experts to create countermeasures, including updates to their Misrepresentation policy aimed at suspending advertisers promoting these scams. As a result, more than 700,000 fraudulent advertiser accounts were permanently suspended, leading to a 90% decrease in reports of such scam ads last year.
While these statistics are significant, Google and other major tech companies, like Meta, continue to be primary sources of scam accounts online. This has generated repeated requests from financial institutions and fintech firms for social media companies to take responsibility and provide reimbursements to victims of scams originating on their platforms.