Google Gemini Mosilla researchers hacked to hide the phishing prompt

Artificial intelligence is everywhere these days – in your phone, your car, even your washing machine. I saw the specialty of the underlying AI in another day. And while it may look a bit at the top, no one denys that artificial intelligence has made life easier in many ways.

From increasing productivity to unlocking new creative equipment, it is changing how we work and live. The most common version that you probably have faced? Think about the chatbots like Generative AI, chat. But this technique can be as helpful, it is not without its problems.

If you have used Google’s workspace suit, you must have seen the company’s AI model, Gemini, which is integrated in apps such as Docs, Sheets and Gmail. Now, researchers say that the attackers can secretly manipulate the Gemini-based email summary in hidden fishing signals.

Sign up for my free cyber report
Take my best technical tips, immediate safety alerts and exclusive deals directly to your inbox. In addition, you will get immediate access to my final scam survival guide – when you join me Cyberguy.com/newsletter

How AI is now helping hackers fool your browser’s safety equipment

How can Gemini Summary be hacked

Researchers at 0DIN’s 0DIN have discovered a vulnerability in the Gemini of Google for the scope that allows the attackers to inject the instructions hidden in the email summary. The issue displayed by Marco Figuro shows how the generative AI tool can be misled through indirect early injections. This technique embedded the invisible command inside an email body. When Gemini summarizes the message, it interprets and acts on those hidden signals.

What is Artificial Intelligence (AI)?

The attack does not depend on the suspected link or attachment. Instead, it uses a combination of HTML and CSS to hide the prompt by setting the font shape to zero and color to white. These commands are invisible in the standard view of Gmail, but are still accessible to Gemini. Once you request a summary, AI can be cheated in fake safety alerts coming from Google or immediate instructions.

In a proof of the concept, Mithun warned a user that a Gmail password was compromised and included a fake support phone number. Since the Gemini summary is integrated into the Google work area, you are more likely to rely on the information, making this strategy particularly effective.

What is Google doing about defects?

While Google has implemented rescue against early injections since 2024, this method bypasses the current security. The company told Cybergui that it is actively deploying the updated security measures.

Get Fox Business when you click here

In a statement, a Google spokesman said, “Avoidance against industry affected attacks, such as early injections, have been a continuous priority for us, and we have deployed several strong defense to protect users, including security measures to prevent harmful or misleading reactions.

Google also confirmed that it had not seen active exploitation of this specific technique.

6 ways you can be safe from AI Fishing Scam

So, how can you protect yourself from fishing scams that exploit AI devices like Gemini? Here are six essential steps that you can take to stay safe right now:

1. Do not rely on AI-related materials

Just because a summary appears in Gmail or docks do not mean that it is automatically safe. AI- Treat with the same cautious that you will do any unwanted message with the same cautious suggestions, alerts or links. Always verify important information through official sources, such as security alert or phone number.

2. Avoid using summary facilities for suspicious email

If an email seems unusual, especially if it is unexpected or from someone you do not recognize, avoid using AI summary facility. Instead, read the entire email as it was originally written. This reduces the possibility of falling for misleading summary.

3. Beware of fishing emails and messages

Look for emails or messages that create a sense of urgency, ask you to verify the account details or provide unexpected links or contact information, whether they appear reliable or come from familiar sources. The attackers can use AI to crafts requests for realistic looking alerts or sensitive information, sometimes hidden within the summary generated automatically. Therefore, always stop and investigate suspicious signals before responding.

The best way to protect yourself from malicious links that establishes malware, potentially reaches your personal information, is to install antivirus software on all your devices. This security can also make you alert for email and ransomware scams, keeping your personal information and digital assets safe.

Get my pics for the best 2025 antivirus protection winners for your Windows, Mac, Android and iOS devices Cyberguy.com/Lockupyourtch

4. Keep your apps and extensions updated

Ensure that Google workpiece and your browser are always running the latest version. Google regularly releases security updates that help prevent new types of attacks. In addition, avoid using informal extensions that have access to your gmail or docks.

5. Invest in data removal service

Gemini summary attacks such as AI-operated scams do not occur in a vacuum. They often start with personal information of theft. This data can come from previous violations, public records or details that you have inadvertently shared online. The service to remove and request a data removal by removing and requesting your information from the data broker sites may continuously scan and request. While no service can erase everything, reducing your digital footprint makes it difficult to personalize fishing efforts for the attackers or link you to the known breech data. Think of it as another layer of security in a world where AI makes target scams even easier.

See my top pics for data removal services and get a free scan to find out if your personal information is already on the web Cyberguy.com/Delete

Get a free scan to know if your personal information is already on the web: Cyberguy.com/freescan

6. If you are worried then disable AI summary for now

If you are worried about falling for AI-borne fishing effort, consider disabling Gemini summary in Gmail until Google leaves strong security. You can still read full email in a traditional way, which can reduce your risk of misleading from your summary.

How to disable Gemini features on desktop

• open Gmail on the desktop.
• Click setting Gear icon on the upper right.
• Click See all settings,
• Scroll “Google Workpagged Smart Facilities” And click Manage the workspace smart feature settings,
• Disabled Togal for Smart features in google field,
• Then, click save.
• Comment: Will do this Stop Gemini Summary Also other smart features.

How to disable Gemini features on mobile

On iphone:

If you especially use Mithun Mobile App:

• Open Gemini app.
• Tap yourself profile picture.
• Tap Mithun Apps Activity.
• At the top, tap Close

On Android:

SEttings may vary depending on the manufacturer of your Android phone

• Open the gmail app On your android.
• Tap Menu In icons (three horizontal lines) upper left corner.
• Scroll down and tap setting,
• Select Relevant email account.
• Scroll down and tap Google Western Smart Facilities And Uncheck the box To close them.

Major warnings to learn:

• Disable Smart features Other convenient functionalities can be removed, such as future text and automated appointment detection.
• Gemini icon or summary button may still be visibleEven after disabled these features. Some users report physically hiding them through the browser tools.

There is no centralized single “off switch” to completely remove all Gemini AI references everywhere, but these phases significantly reduce the appearance and risk of the feature.

Click here to get Fox News app

Kurt’s key to Techway

This vulnerability throws light on how the phishing strategy is developing along with AI. Instead of relying on visible red flags such as Mrs.Lade URL or suspected attachment, the attackers are now targeting reliable systems that help users to filter and explain messages. Since the AI productivity becomes more deeply embedded in devices, early injections can emerge as a subtle but powerful vector for social engineering, which hides malicious intentions in many devices designed to simplify communication.

How ease you are giving AI a brief or filtering your email, and where do you pull the line? Write us and tell us Cyberguy.com/Contact

Sign up for my free cyber report
Take my best technical tips, immediate safety alerts and exclusive deals directly to your inbox. In addition, you will get immediate access to my final scam survival guide – when you join me Cyberguy.com/newsletter

Copyright 2025 cyberguy.com. All rights reserved.