AI Content Hijacks Google Discover to Deliver Malicious Alerts

A new large-scale cyber operation is exploiting Google’s Discovery feed to spread malicious notifications and scams through AI-generated content.

Pushpaganda begins with threat actors creating around 113 fake domains filled with AI-written articles and clickbait headlines.

These posts are crafted to appear in Google’s Discovery feed the stream of news and personalized content visible on Android home screens and Chrome’s new tab pages.

Using advanced SEO techniques and possible paid placements, attackers inject these deceptive stories directly into users’ content streams.

The articles often feature attention-grabbing headlines about taxes, currency benefits, or too-good-to-be-true tech deals such as “$1390 IRS Deposit Approved” or claims of ultra-cheap smartphones.

Researchers from HUMAN’s Satori Threat Intelligence and Research Team have uncovered this sophisticated threat, dubbed Pushpaganda, which combines ad fraud, social engineering, and scareware tactics to manipulate mobile users worldwide.

When users click these links, they are redirected to an attacker-controlled website, prompting them to enable push notifications.

Once permission is granted, the attackers gain the ability to deliver persistent, system-level notifications that bypass traditional ad blockers.

Scareware and Social Engineering at Scale

The notifications sent through Pushpaganda are designed to alarm or deceive. Victims receive fake legal threats, fabricated bank notices, or impersonated messages such as missed calls from family members.

Clicking any of these messages leads to another Pushpaganda domain, creating a continuous loop of deceptive content and fake alerts.

Researchers are unsure if Pushpaganda-related content showing in users’ Discovery feeds is purchased by the threat actors or if there are advanced SEO techniques at play.

At its height, HUMAN recorded over 240 million ad bid requests in a single week linked to this operation. Initially focused on Indian users, the campaign has since expanded across the United States, Australia, and other regions.

Each interaction ad impressions, clicks, or page views is fraudulent because it does not represent genuine user interest but scripted engagement driven by manipulation.

In addition to push notification abuse, some Pushpaganda pages include deceptive buttons labeled “Apply Now” or “Claim Offer.”

These buttons don’t lead to legitimate destinations; instead they execute JavaScript redirections that open new browser tabs or silently reload other actor-controlled pages in the background.

This rotation of inactive tabs inflates traffic metrics, creating the illusion of authentic user engagement while generating additional ad revenue.

Google Acts, HUMAN Defends

Satori researchers also discovered the use of deepfake advertisements embedded throughout these domains.

Satori researchers identified a JavaScript rotation algorithm that forces inactive tabs or background windows to automatically cycle through various actor-owned pages.

These fabricated images and videos, often depicting celebrities or medical experts, are used to bolster credibility and lure victims into further scams or product purchases.

After receiving the list of associated domains from the Satori team, Google confirmed that it has implemented a fix to block low-quality, manipulative content like Pushpaganda’s from appearing in Discovery feeds.

Satori continues to monitor for new domains or evolutions of the Pushpaganda infrastructure. The campaign’s abuse of AI-generated content and trusted mobile interfaces demonstrates how threat actors are increasingly merging advertising manipulation with modern AI tools to deceive users at massive scale.

Meanwhile, customers using HUMAN’s Ad Fraud Defense and Ad Click Defense solutions are already protected from these fraudulent activities.

Follow us on Google News, LinkedIn, and X to Get Instant Updates and Set GBH as a Preferred Source in Google.