started · updated
Google deceptive software ads reached over 100 million impressions
A joint study by New York University Tandon and Radboud University has revealed that deceptive software advertisements have bypassed Google’s moderation systems, accumulating over 100 million impressions. Using an AI tool named AdLens, researchers analyzed approximately 188,834 software-related ads from Google’s Ads Transparency Center recorded through March 2026.
The study identified three primary categories of violating content: scareware, which uses fake virus or security alerts to intimidate users; false promises, such as claims of recovering deleted photos or tracking GPS locations via phone numbers; and identity concealment, where advertisers hide their true intent behind vague buttons. Scareware ads alone accounted for over 16 million impressions, while ads making false promises reached approximately 89.9 million impressions.
Researchers also tested Google’s official reporting mechanism, finding that out of 24 suspicious ads reported, only six were removed, while 15 were deemed non-violating by the platform. The report highlights a systemic issue where removing a single ad often fails to stop broader malicious networks, noting that 41 ads linked to a single known malware site remained active even after related content was taken down.
Entities
AdLens · Google · New York University Tandon School of Engineering · Radboud University