Google Warns Users About New AI Scam Targeting Accounts
Google has alerted its 1.8 billion users about a fresh scam that exploits artificial intelligence technology used by cyber criminals. According to tech expert Scott Polderman, the scam employs Gemini, Google’s AI chatbot, to trick users into revealing their passwords.
Polderman noted that hackers are now using Gemini against its own users by sending emails with hidden messages. These messages can prompt Gemini to disclose sensitive login information without the user even being aware of it. “This scam is notable because it’s essentially AI being turned against itself,” he explained.
What makes this scam particularly concerning is that it doesn’t require users to click on any links. Instead, Gemini itself alerts users about potential threats, which can make the scam more deceptive. Google has reiterated that it will never ask users for their login credentials or issue alerts through Gemini.
Another tech expert, Marco Figueroa, highlighted the method behind the scam. Hackers send emails with invisible prompts that only Gemini can detect. These prompts use a zero font size and white text to evade detection by the recipient.
Some users have suggested ways to protect themselves, like disabling certain features in Gmail and adjusting Google Workspace settings to stop Gemini’s functionality. Individuals are encouraged to turn off “SMART FEATURES” and, if necessary, disable the Gemini app completely.
There is growing frustration among users, with some expressing their desire to revert to simpler methods of communication, like pen and paper. “I’m tired of this! I think I’ll change my password just in case,” remarked one user, while another said they had stopped using Gmail altogether after hearing about the scam.
In its security blog, Google noted the rising threats tied to new AI technologies. They explained that attackers are now using a sophisticated technique known as “indirect prompt injections,” where harmful instructions are concealed in everyday communications such as emails or calendar invites.
To help address these risks, Google is implementing layered security measures. They are enhancing the Gemini system itself and developing machine learning models to identify and combat malicious attempts effectively. Their aim is to make it more challenging and costly for hackers to successfully execute such scams.
The message to users is clear: Stay cautious and keep your information secure.
