Whether it’s a fake miracle cure advertisement or a video featuring a celebrity’s cloned voice, deceptive online content has become increasingly common—and artificial intelligence is making it worse.
The widespread availability of AI technology has intensified the longstanding problem of digital fraud and spam that has plagued the internet since its early days. However, as criminals gain access to these powerful tools, major technology companies are also strengthening their own AI defenses to combat the growing threat.
“It’s not that this is a new problem. It is an old problem, supercharged,” said Nate Elliott, a principal analyst at Emarketer. “The biggest difference is the speed and the scale that AI offers both the good actors and the bad actors.”
In its annual advertising safety report released Thursday, Google acknowledged that fraudsters are launching increasingly complex malicious campaigns while highlighting how its AI-driven systems serve as powerful protection tools.
The tech giant’s AI platform called Gemini successfully identified and stopped more than 99% of rule-breaking advertisements before users ever saw them during the past year.
Throughout 2025, Google eliminated or blocked over 8.3 billion advertisements, which included 602 million ads containing violations typically linked to fraudulent schemes. The company also shut down more than 4 million advertiser accounts due to scam-related activities.
Google maintains its position as a leading player in online advertising. According to Emarketer data, the company generated over $200 billion in global advertising revenue last year, though researchers predict Meta will surpass Google’s performance by 2026.
The company employs thousands of workers dedicated to developing and implementing its advertising standards on a massive scale. Keerat Sharma, Google’s vice president and general manager of ads privacy and safety, explained that incorporating advanced AI into their protection framework has produced more effective outcomes against harmful content.
Gemini now enables the team to examine hundreds of billions of data points—such as account history, user behavior patterns, and advertising campaign characteristics—to better understand the “nuance of what an advertisers intent actually is,” Sharma said. This capability allows them to accurately assess legitimacy or identify potentially harmful advertiser motives. Achieving this sophisticated analysis has also helped legitimate businesses keep their advertisements active, with the report showing that wrongful advertiser account suspensions decreased by 80% last year.
Gemini has also dramatically improved processing speed, according to Sharma. Previously, examining digital elements within an advertisement could require several seconds to minutes or longer, but now this analysis occurs within milliseconds. This speed “allows us to stop things right at the front door,” he said. Google also utilizes various other protective measures, including a comprehensive advertiser verification system, that work collectively to strengthen security.
The types of content Google targets for removal span a wide range. Problematic advertisements could appear as “all the forms of spam and scam that have always existed, just people are able to produce them faster and at higher volume,” Elliott said.
Specialists who discussed the issue with The Associated Press indicated that the ongoing battle between AI-driven fraud and AI-powered protection systems will continue as technology evolves.
“We’re already close, but it’s going to be heading even more to (where) it’s just AI versus AI,” said Matt Seitz, the director of the AI Hub at the University of Wisconsin-Madison. “The volume of this problem is so large that it can’t be managed directly through humans.”
