The stolen data includes names, emails, limited billing information, and some government-ID images.

Popular social platform Discord has suffered a data breach—though technically, it wasn’t Discord itself that was hacked. A third-party customer support provider was compromised, allowing attackers to access Discord’s user data. Either way, it’s Discord users who feel the impact.

The breach, which happened on September 20, didn’t involve a direct attack on Discord’s servers. Instead, attackers gained access through a customer support partner. Criminals claimed they breached Zendesk, the help desk service Discord uses for customer support, according to reports.

Attackers stole data including real names, Discord usernames, email addresses, and other contact details provided to customer support. The breach appears to have been financially motivated and included a ransom demand.

In some cases, “limited billing information” was also taken—including payment type, the last four digits of credit card numbers, and purchase histories. Customer IP addresses and messages with support agents were also exposed.

More concerning is that some users had especially sensitive information stolen. Discord said in its advisory:

> “The unauthorized party also gained access to a small number of government-ID images (e.g., driver’s license, passport) from users who had appealed an age determination. If your ID may have been accessed, that will be specified in the email you receive.”

Attacks like this show how large the fallout can be when consumer-focused services are hit.

Discord, once known mainly for gaming communities, now hosts more than 200 million monthly active users and is widely used by companies to host customer and community channels.

According to vendor risk management firm Rescana, the attackers identified themselves as Scattered Lapsu$ Hunters (SLH). BleepingComputer reported this too, but later said SLH changed its story, pointing to another group it knows and interacts with.

The problem with these kinds of groups is that they often share techniques and even members, muddying the waters.

Rescana described SLH as a coalition—combining tactics from Scattered Spider, Lapsu$, and ShinyHunters: groups known for stealing data from third-party partners like support vendors or software suppliers. The attacks relied on social engineering rather than malware.

Discord disclosed the incident 13 days later, on October 3. It has since revoked its support provider’s access, launched an internal investigation with a forensics firm, and notified affected users.

The company reminded users that any communication about the breach will come only from noreply@discord.com and that it will never call users directly.

**We don’t just report on data privacy—we help you remove your personal information**

Cybersecurity risks should never spread beyond a headline. With Malwarebytes Personal Data Remover, you can scan to find out which sites are exposing your personal information, and then delete that sensitive data from the internet.

 Discord warns users after data stolen in third-party breach 

Attackers are using realistic-looking 1Password emails to trick users into handing over their vault logins.

In a very recent and well-targeted phishing attempt, scammers tried to get hold of the 1Password credentials belonging to a Malwarebytes’ employee.

Stealing someone’s 1Password login would be like hitting the jackpot for cybercriminals, because they potentially export all the saved logins the target stored in the password manager.

The phishing email looked like this:

> “Your 1Password account has been compromised
> 
> Unfortunately, Watchtower has detected that your 1Password account password has been found in a data breach. This password protects access to your entire vault.
> 
> Take action immediately
> 
> To keep your account secure, please take the following actions:
> 
> – Change your 1Password account password
> 
> – Enable two-factor authentication
> 
> – Review your account activity
> 
> Secure my account now
> 
> If you need help securing your account, or have any questions, contact us. Our team is on hand to provide expert, one-on-one support.”

While the email looks convincing enough, you can spot a few red flags.

*   The sender’s address watchtower@eightninety[.]com does not belong to 1Password, which typically use the domain @1password.com.
*   If you hover over the “Secure my account now” button you’ll notice that it points to: https://mandrillapp[.]com/track/click/30140187/onepass-word[.]com?p={long-identifier}

Although 1Password’s Watchtower feature can send alerts about compromised passwords, it does so by checking its database of known data breaches and then notifying you directly within the 1Password app or through very specific emails about the breach—not by sending a generic message like this.

Obviously, the onepass-word[.]com is a feeble attempt to make it look legitimate. I guess all the good typosquats were already taken or protected. What’s interesting is that the “Contact us” link goes to the legitimate support.1password.com, although it also flows through a redirect through mandrillapp.

Mandrillapp is a transactional email API and delivery service provided by Mailchimp. It enables organizations to send automated, event-driven emails like order confirmations, password resets, and shipping notifications. Mandrill also provides delivery tracking and statistics to their customers.

What the scammers may not have realized is that Mandrillapp doesn’t forward people to known phishing websites.

Shortly after the emails went out on October 2, the domain was already classified as a phishing site by several vendors. By October 3, anyone that clicked the button would end up viewing an error message on mandrillapp[.]com saying bad url - reference number: {23 character string}.

But early birds would have seen this form:

Anyone who fell for this scam would have sent their 1Password credentials straight to the phishing crew.

On September 25, 2025, Hoax-Slayer reported about a very similar phishing expedition. This might indicate that this was the first—and probably is not the last—attempt, so be warned.

With the key to your password vault, cybercriminals could take over all your important accounts and potentially steal your identity, so be very careful about where and when you use these credentials.

****Our advice:****

*   **Do not** click any links or buttons in an unsolicited email
*   **Do not** provide any of your 1Password credentials or personal information.
*   If you are concerned about your 1Password account, go directly to the official 1Password website or app and check your account status there.
*   Use up-to-date real-time protection which includes a web protection module.

**Indicators of compromise (IOCs)**

Email address:

watchtower@eightninety[.]com

Domain Phishing website:

onepass-word[.]com

**We don’t just report on threats – we help safeguard your entire digital identity**

Cybersecurity risks should never spread beyond a headline. Protect your—and your family’s—personal information by using identity protection.

 Phishers target 1Password users with convincing fake breach alert 

This week on the Lock and Code podcast, we speak with Twitter's first employee, Rabble, about saving the best parts of social media today.

_This week on the Lock and Code podcast…_

“Connection” was the promise—and goal—of much of the early internet. No longer would people be separated from vital resources and news that was either too hard to reach or made simply inaccessible by governments. No longer would education be guarded behind walls both physical and paid. And no longer would your birthplace determine so much about the path of your life, as the internet could connect people to places, ideas, businesses, collaborations, and agency.

Somewhere along the line though, “connection” got co-opted. The same platforms that brought billions of people together—including Facebook, Twitter, Instagram, TikTok, and Snapchat—started to divide them for profit. These companies made more money by showing people whatever was most likely to keep them online, even if it upset them. More time spent on the platfrom meant more likelihood of encountering ads which meant more advertising revenue for Big Tech.

Today, these same platforms are now symbols of some of the worst aspects of being online. Nation-states have abused the platforms to push disinformation campaigns. An impossible sense of scale allows gore and porn and hate speech to slip by even the best efforts at content moderation. And children can be exposed to bullying, peer pressure, and harassment.

So, what would it take to make online connection a good thing?

Today, on the Lock and Code podcast with host David Ruiz, we speak with Rabble—an early architect of social media, Twitter’s first employee, and host of the podcast Revolution.Social—about what good remains inside social media and what steps are being taken to preserve it.

> “ I don’t think that what we’re seeing with social media is so much a set of new things that are disasters that are rising up from this Pandora’s box… but rather they’re all things that existed in society and now they’re not all kept locked away. So we can see them and we have to address them now.”

Tune in today to listen to the full conversation.

**Listen up—Malwarebytes doesn’t just talk cybersecurity, we provide it.**

Protect yourself from online attacks that threaten your identity, your files, your system, and your financial well-being with our exclusive offer for Malwarebytes Premium Security for Lock and Code listeners.

 What&#8217;s there to save about social media? (Lock and Code S06E20) 

Step-by-step instructions on how to enable 2FA on your Facebook account—for Android, iOS, and via the website.

While two-factor authentication (2FA) is not completely fool-proof, it is one of the best ways to protect your accounts from hackers. It adds an extra step when logging in, which is a small extra effort for you, but it dramatically boosts your security.

With 2FA, you’ll be asked for a special login code when signing in from a device or browser Facebook doesn’t recognize—even if someone already knows your password.

Here’s how to enable 2FA on Facebook for Android, iOS, and the web.

****How to set up 2FA for Facebook on Android****

1.  Open the **Facebook app** (make sure you’re signed in).
2.  Tap the **menu** (three horizontal lines).
3.  Choose **Settings & Privac**y > **Settings**.
4.  In the **Accounts Center** tap **Password and security**.
5.  Tap **Two-factor authentication** and select your account your want to protect.
6.  Re-enter your password. Facebook will send a one-time code to your phone or email to confirm it’s you.
7.  Pick your preferred security method:
    *   **Authentication app (recommended)** – such as Google Authenticator or Authy.
    *   **Text message (SMS) or WhatsApp** – codes sent to your phone number.
    *   **Security key** – a USB or Bluetooth device.
    *   **Recovery codes** – backup codes to use if other methods aren’t available.
8.  Follow on-screen instructions to complete the setup.

**How to set up 2FA for Facebook on iPhone or iPad**

1.  Open the **Facebook app** (make sure you’re signed in).
2.  Tap your **profile picture** in the bottom right corner.
3.  Go to **Settings & Privacy > Settings**.
4.  Tap on **Accounts Center**, then **Password and security**.
5.  Tap **Two-factor authentication** and select your account.
6.  Re-enter your password. Facebook will send a one-time code to your phone or email to confirm your identity.
7.  Choose your preferred method:
    *   **Authentication app (recommended)** – such as Google Authenticator or Authy.
    *   **Text message (SMS) or WhatsApp** – codes sent to your phone number.
    *   **Security key** – a USB or Bluetooth device.
    *   **Recovery codes** – backup codes to use if other methods aren’t available.
8.  Follow on-screen instructions to complete the setup.

**How to set up 2FA for Facebook on the web**

1.  Go to facebook.com/settings (or from the home screen, click your profile picture and then **Settings & privacy**).
2.  Navigate to **Password and security**.  
    
3.  Click **Two-factor authentication**, then select your account.
4.  Facebook will send a one-time code to your WhatsApp or email to confirm it’s you, and may ask you to re-enter your password.
5.  Choose your preferred method:
    *   **Authentication app (recommended)** – such as Google Authenticator or Authy.
    *   **Text message (SMS) or WhatsApp** – codes sent to your phone number.
    *   **Security key** – a USB or Bluetooth device.
    *   **Recovery codes** – backup codes to use if other methods aren’t available.
6.  Follow on-screen instructions to complete the setup.

**Why you should **enable it today****

Even the strongest password can be stolen. With 2FA, attackers would also need access to your additional factor to be able to log in to your account, whether that’s a code on a physical device or a security key. That makes hijacking your account much harder.

We recommend you set up 2FA on all your important accounts, including messaging and social media accounts. It only takes a few minutes, but can save you from hours or even days of stress later. It’s currently the best password advice we have.

**We don’t just report on threats – we help protect your social media**

Cybersecurity risks should never spread beyond a headline. Protect your social media accounts by using Malwarebytes Identity Theft Protection.

 How to set up two-factor authentication (2FA) on your Facebook account 

A list of topics we covered in the week of September 29 to October 5 of 2025

October 3, 2025 - After posting children’s photos online and issuing ransom demands, cybercriminals targeting Kido nurseries say they’ve erased the stolen data.

October 2, 2025 - Meta has announced it will start using your interactions with its generative AI to serve targeted ads.

October 2, 2025 - Cybercriminals are targeting older Facebook users with fake community and travel groups that push malicious Android apps.

October 1, 2025 - Google’s Gemini AI suite had vulnerabilities that let attackers hide malicious instructions in everyday web activity.

September 30, 2025 - Researchers found several security problems in Life360's Tile trackers, most of which could be solved with encryption.

 A week in security (September 29 &#8211; October 5) 

After posting children’s photos online and issuing ransom demands, cybercriminals targeting Kido nurseries say they’ve erased the stolen data.

Last week we yelled at some “hackers” that threatened parents after stealing data from their children’s nursery.

This followed a BBC report that a group calling itself “Radiant” claimed to have stolen sensitive data related to around 8,000 children from nursery chain Kido, which operates in the UK, US, China, and India.

To prove their possession of the data, the criminals posted samples on their darknet website, including pictures and profiles of ten children. They then issued a ransom demand to Kido, threatening to release more sensitive data unless they were paid.

A few days later, they added profiles of another ten children and threatened to keep going until Kido paid their ransom demand. The group also published the private data of dozens of employees including names, addresses, National Insurance numbers, and contact details.

The criminals then reportedly contacted parents directly with threatening phone calls whilst pushing to get their ransom paid.

But after massive pushback from the general public and some prominent members of the malware community, the attackers initially blurred the children’s images but left the data online. Soon after, they pulled everything offline and issued an apology.

They even claim to have deleted all the children’s data. One of the cybercriminals told the BBC:

> “All child data is now being deleted. No more remains and this can comfort parents.”

But, as we have mentioned many times before, computers—and the internet in particular—are not very good at “forgetting” things. Data tends to pop up in unexpected places. Remember when supposedly deleted iPhone photos showed up again after an iOS update?

And, of course, all we have to go on is the word of a criminal with such a bad reputation that even they seemed ashamed of what they did.

They might be feeling a bit sorry for themselves, as they claim to have paid an initial access broker (IAB) for the access to Kido’s systems and will likely see no return on that “investment”.

**Protecting yourself after a data breach**

There are some actions you can take if you are, or suspect you may have been, the victim of a data breach.

*   **Check the vendor’s advice.** Every breach is different, so check with the vendor to find out what’s happened and follow any specific advice it offers.
*   **Change your password.** You can make a stolen password useless to thieves by changing it. Choose a strong password that you don’t use for anything else. Better yet, let a password manager choose one for you.
*   **Enable two-factor authentication (2FA****).** If you can, use a FIDO2-compliant hardware key, laptop or phone as your second factor. Some forms of 2FA can be phished just as easily as a password, but 2FA that relies on a FIDO2 device can’t be phished.
*   **Watch out for fake vendors.** The thieves may contact you posing as the vendor. Check the company’s website to see if it’s contacting victims and verify the identity of anyone who contacts you using a different communication channel.
*   **Take your time.** Phishing attacks often impersonate people or brands you know, and use themes that require urgent attention, such as missed deliveries, account suspensions, and security alerts.
*   **Consider not storing your card details**. It’s definitely more convenient to get sites to remember your card details for you, but we highly recommend not storing that information on websites.
*   **Set up identity monitoring**, which alerts you if your personal information is found being traded illegally online and helps you recover after.

**We don’t just report on data privacy—we help you remove your personal information**

Cybersecurity risks should never spread beyond a headline. With Malwarebytes Personal Data Remover, you can scan to find out which sites are exposing your personal information, and then delete that sensitive data from the internet.

 From threats to apology, hackers pull child data offline after public backlash 

Meta has announced it will start using your interactions with its generative AI to serve targeted ads.

Meta has announced that conversations with its AI assistant will soon be used for targeted advertising. If you’re the kind of person that notices ads for products just after you spoke about them, you won’t be happy about this update.

Meta AI is the company’s generative AI assistant, built into Facebook, Instagram, WhatsApp, Messenger, and Threads. It can answer questions, generate text or images, and recommend content.

Users will soon start to receive notifications about how their interactions with Meta’s generative AI features will be used for targeted advertising. So, ask Meta AI about vacations, hobbies, or new gadgets, for example, and you might soon find related ads in your feed.

Certain topics are excluded—religious views, sexual orientation, political views, health, racial or ethnic origin, philosophical beliefs, or trade union membership—but everything else is fair game.

Meta said this update takes effect on December 16, 2025, and will start notifying users on October 7, 2025 through in-product notifications and emails.

Thanks to stricter privacy laws, users in the EU, UK, and South Korea are exempt, The Register reports.

According to Meta, over 1 billion people use its AI every month. And as we all know, targeted ads bring in more money than generic ones. So, this is how Meta plans to earn back all the money it spent on AI development.

Because, like it or not, Meta isn’t really about connecting friends all over the world. Its business model is almost entirely based on selling targeted advertising space across its platforms.

Generative AI providers are increasingly weaving advertising into their products, especially in free or freemium offerings. Many companies now use AI to create personalized ads directly within user interactions. For example, AI-powered recommendation engines analyze user data and behavior to deliver highly targeted ads, boosting relevancy and engagement. Done well, this approach makes ads feel less intrusive and more like natural content suggestions tailored to individual preferences.

Still, the industry faces big ethical and privacy challenges. Brands and AI providers must balance personalization with transparency and user control, especially as AI tools collect and analyze sensitive behavioral data. Many are turning to opt-in mechanisms, clearer privacy settings, and responsible data use policies to maintain user trust while taking advantage of AI’s ability to deliver relevant, personalized ads.

Meta promises that affected users can continue to adjust the content and ads they’re seeing at any time with tools like Ad Preferences and other feed controls.

The Register jokingly suggested we start our Meta AI chats with something from the “excluded” list, hoping to keep the whole conversation from being used for targeted advertising. Their example:

> “Oh, Lord, Meta really thought this was a good idea?”

In the end, it might be better not to share anything too personal with Meta AI, or any chatbot for that matter, and stick to kittens and puppies instead.

**We don’t just report on threats – we help protect your social media**

Cybersecurity risks should never spread beyond a headline. Protect your social media accounts by using Malwarebytes Identity Theft Protection.

 Your Meta AI conversations may come back as ads in your feed 

Cybercriminals are targeting older Facebook users with fake community and travel groups that push malicious Android apps.

An infostealer and banking Trojan rolled into one is making the rounds in Facebook groups aimed at “active seniors”.

Attackers used social engineering methods to lure targets into joining fake Facebook groups that appeared to promote travel and community activities—such as trips, dance classes, and community gatherings. Once people joined, they were invited to download an Android app to “register” for those offered activities.

Researchers at ThreatFabric found numerous Facebook groups created under this pretense, stocked with AI-generated content to appear authentic and trick users into downloading the malware. App names included Senior Group, Lively Years, ActiveSenior, and DanceWave. In some cases, victims were also asked to pay a sign-up fee on the same website, leading to phishing and card detail theft.

One of the servers hosting these downloads was located at download.seniorgroupapps[.]com.

Sometimes the cybercriminals sent a follow-up message through Messenger or WhatsApp, sharing the download links for the malicious apps.

Often this would be the Datzbro Trojan, but sometimes victims were hit with Zombinder, a Trojan dropper capable of bypassing the security restrictions Google introduced in Android 13 and later versions.

**What Datzbro can do**

The researchers found that Datzbro had capabilities similar to both spyware and banking Trojans—specifically designed to drain bank accounts.

Once installed, this Android malware can:

*   Record audio and video, and access files and photos.
*   Display phishing overlays that mimic other apps to steal passwords and send them to the attackers.
*   Let attackers remotely control infected Android devices, including locking or unlocking the screen.

Researchers analyzed the code and suspect that it was likely developed in China, but later leaked and was reused by broader cybercriminal groups. The campaign has reached victims worldwide, including Australia, Singapore, Malaysia, Canada, South Africa, and the UK.

**How to stay safe in Facebook groups**

Although many of the Facebook groups involved in this campaign have been taken down, there might be others. To protect yourself:

*   Check a Facebook group’s history and avoid those might have freshly set up for malicious purposes. Unfortunately, it’s not possible to check the age of a group before you join, but once you’re a member, look at the dates of historical posts or pinned posts.
*   Don’t click on links or install apps provided by such groups or by private messages from people you don’t really know.
*   Use up-to-date real-time anti-malware protection, especially on your mobile devices.
*   Be wary of groups offering suspicious or too-good-to-be-true promises.
*   Check a group’s description and rules for professionalism or red flags.

It’s worth noting that many of the groups also included a button to download an “iOS application.” These were just placeholders at the time, but might be an indication that there are plans to target iPhone users as well.

**Indicators of Compromise (IOCs)**

**The malicious app used these names:**

Senior Group

Lively Years

ActiveSenior

DanceWave

**and these package names:**

twzlibwr.rlrkvsdw.bcfwgozi

orgLivelyYears.browses646

com.forest481.security

inedpnok.kfxuvnie.mggfqzhl

**We don’t just report on phone security—we provide it**

Cybersecurity risks should never spread beyond a headline. Keep threats off your mobile devices by downloading Malwarebytes for iOS, and Malwarebytes for Android today.

 Scam Facebook groups send malicious Android malware to seniors 

Sendit and its CEO are accused of preying on young users—signing them up illegally, misusing their data, and tricking them with bogus messages and hidden fees.

The Federal Trade Commission (FTC) has sued Sendit’s parent company, saying it signed up children under 13, collected their personal data, and misled them with fake messages and recurring bills.

The lawsuit, filed against the app’s owner Iconic Hearts Holdings Inc and CEO Hunter Rice, alleges the company let users under the age of 13 sign up for Sendit and collected personal information about these users without parental consent—violating the Children’s Online Privacy Protection Rule (COPPA).

Sendit is an add-on for Snapchat and Instagram, rather than a standalone app. Its primary feature is to allow users to post prompts or questions (called a Sendit) on their social media stories and receive anonymous replies from other users.

In 2022, the app registered 116,000 people who self-declared that they were under 13 years old, according to the suit. Even after parents complained, the company continued to collect children’s phone numbers, birthdates, photos, and usernames for Snapchat, Instagram, TikTok and other accounts.

The FTC also alleges that Sendit misled users about its paid “Diamond Membership.” The feature promised to allow users to see who had sent certain messages. In practice, it didn’t reveal the senders, according to the suit. Worse still, the company and its CEO faked some of these messages, the FTC alleges. According to the complaint:

> “Defendants trick users into believing that they have received provocative and sometimes sexual or romantic messages from their social media contacts, when in reality it is often Defendants themselves who have sent those messages.”

Iconic Hearts also failed to disclose recurring charges clearly, according to the FTC—charging up to $9.99 every week after making it look like users were paying a single fee to disclose a user’s identity.

Normally, cases like this end in a settlement. This time, the FTC referred the case to the Department of Justice (DoJ). It does this when it believes that the defendants are violating or about to violate the law, and that referring the case would be in the public interest. So now, the Central District of California will decide the case.

Iconic Hearts also publishes the apps Noteit, Starmatch, and Locksmith. Launched in 2018, Sendit has been downloaded more than five million times on Google Play, and the company claims a total user base of around 25 million. The company has claimed Sendit is “the top Gen Alpha social networking app.”

This isn’t the only case where anonymous messaging apps have run afoul of COPPA violations. In July 2024, the FTC settled with NGL Labs and its founders for $5 million. That app was accused of marketing to kids and teens, sending fake messages to drive up usage, tricking users into paid upgrades, and sneaking in recurring charges.

“Company executives told employees to reach out to high school kids directly,” said the FTC at the time. NGL Labs also falsely claimed that AI content moderation filtered harmful messages like cyber bullying, the Commission added. The settlement banned NGL from marketing its app to anyone under 18.

What could this mean for Iconic Hearts? The current maximum penalty enforceable by courts for failing to comply with COPPA is $53,088 per violation, according to the FTC.

DoJ COPPA-related suits on the FTC’s behalf are not unheard of. Epic Games got a record $275 million penalty for COPPA violations in December 2022 after the DOJ sued it on behalf of the FTC (alongside another $245 million penalty for using ‘dark patterns’ to mislead users).

Epic Games was aware that many children were playing its Fortnite game, yet it collected personal data from children without first obtaining parents’ verifiable consent, the suit said. The company also made it difficult for parents to delete their children’s personal information, and sometimes didn’t do as asked.

The takeaway from this story? Try to keep kids under 13 off social media apps as long as possible, and when the time does come, stay involved. Talk to them about online safety, monitor their usage, and keep the conversation open.

**We don’t just report on threats – we help protect your social media**

Cybersecurity risks should never spread beyond a headline. Protect your social media accounts by using Malwarebytes Identity Theft Protection.

 Sendit tricked kids, harvested their data, and faked messages, FTC claims 

Google’s Gemini AI suite had vulnerabilities that let attackers hide malicious instructions in everyday web activity.

Security researchers discovered three vulnerabilities in Google’s Gemini artificial intelligence (AI) assistant. Although now patched, this “Trifecta”, as the researchers called it, raises important questions about how safe AI tools really are, especially as they become a part of services many of us use on a daily basis.

The flaws were found in three different Gemini components:

*   **Gemini Cloud Assist**, which summarizes logs for cloud services, could be tricked by hidden prompts inside web requests. Attackers could exploit this flaw to sneak malicious instructions into the system, potentially gaining control over cloud resources.
*   **Gemini Search Personalization Model** could inject harmful prompts into a user’s Chrome browsing history by getting them to visit a special website. If the user later interacted with Gemini’s personalized search AI, the injected commands could force the AI leak to personal data, including saved information and location.
*   **Gemini Browsing Tool** could be tricked into sending stored user information and location data to a malicious server through its web page summarization feature.

Google fixed these issues by blocking Gemini from rendering dangerous links and strengthening its defenses against such prompt injections. But if you used Google services that rely on Gemini AI, there is a chance these vulnerabilities were exploited before the patch—especially if you visited a malicious website or used Gemini features tied to cloud services.

These vulnerabilities are prime examples of how AI, despite its benefits, can open new attack avenues. Attackers may hide malicious instructions inside ordinary files and web requests, fooling AI into performing harmful actions without any obvious warning signs.

For everyday users, the risk is low—Google has already patched these vulnerabilities. But this news reminds all of us that AI security is an evolving concern, especially as new features and use-cases may be developed with security as an afterthought.

**How to safely use AI**

These flaws show that AI systems themselves can be used as a method for attacks, not just a target. This is important as AI becomes more embedded in cloud services and applications.

You should be cautious about:

*   Avoid visiting unknown or suspicious websites, especially those that prompt you to interact with AI assistants.
*   Keeping software, browsers, and apps up to date to benefit from security patches.
*   Be mindful of the information you share with AI tools.
*   Use a real-time anti-malware solution, preferably with web protection.

**We don’t just report on threats—we remove them**

Cybersecurity risks should never spread beyond a headline. Keep threats off your devices by downloading Malwarebytes today.

Source

Malwarebytes