Google Gemini Exploited: AI Attack Targets 1.8 Billion Gmail Users

Abdullah Mustapha

July 17, 2025

Over 1.8 billion Gmail users are open to a new form of cyberattack. The threat uses artificial intelligence in tricking individuals into handing over their own data, such as passwords. It is an emerging problem, experts add—and the majority of users have no idea that it is occurring.

1.8 Billion Gmail Users Put in Urgent Alert: AI Attack Swiping Passwords in Stealth

The plan aims at Google’s Gemini AI, an inbuilt feature in Gmail and other Google services. Attackers are embedding hidden commands within emails. If a user clicks on “summarize this email,” Gemini captures the hidden instructions instead of the text that is seen. This results in the AI creating deceptive security warnings or critical messages.

These spams may claim that hackers have compromised your Gmail account. They can even provide a fake “Google support” phone number and ask you to call in for help. But the phone number goes to the scammers, not Google.

This is how it works: the hackers use white text on a white background and make the font size zero. This conceals the malicious commands from human sight—but not from AI tools like Gemini. This is a “trick injection” of sorts, which “fools” the AI into answering the hidden message.

Security researchers, including Mozilla’s 0Din security team, discovered this stunt. They illustrated how Gemini is manipulated into displaying spurious warnings that mimic authentic ones. These warnings then prompt users to perform an action that leads to phishing sites or scam calls.

To protect yourself, professionals offer basic guidelines:

Never trust security alerts from Gemini summaries. Google doesn’t use Gemini for genuine warnings.
Watch out for emails that demand you act straight away, especially those asking you to click links or telephone numbers.
Email programs can be configured to ban or mark hidden content like zero-size fonts and white-on-white text.
Set up filters to detect unusual patterns like many urgent messages, weird URLs, or odd telephone numbers.

Until Google remediates this vulnerability, attackers will keep exploiting it. Since Gemini and other AI software are integrated into Google Docs, Calendar, and third-party apps as well, the threat is spreading.

Beware. Delete any suspicious summaries. And always verify alerts via official Google media, not AI-generated text.

Disclaimer: We may be compensated by some of the companies whose products we talk about, but our articles and reviews are always our honest opinions. For more details, you can check out our editorial guidelines and learn about how we use affiliate links.

Follow Gizchina.com on Google News for news and updates in the technology sector.

Source/VIA :

TheSun