Insights / Health and Life Sciences

Cybersecurity and AI: Cybercriminals use ChatGPT to spread malware

Fake ChatGPT websites and apps are already being used to steal data. Learn how cybercriminals exploit AI's popularity to spread malware, and how to stay protected.

Team Stefanini ・ 6 Jan. 2026 ・ Estimated read: 4 min

Experts say that while there is a possibility that intelligent chatbots like ChatGPT could be used in the future to generate malicious code, the current focus is on the tool’s popularity. It is believed that fraudsters have increasingly taken advantage of fake ChatGPT websites and apps to lure unwary victims and steal sensitive information.

Meta warns of malware disguised as ChatGPT

On Wednesday, researchers at Meta — Facebook’s parent company — issued a troubling warning: Malicious groups like Ducktail and NodeStealer are impersonating ChatGPT and other similar tools to infect people with malicious browser extensions, ads, and other methods on various social media platforms. All of this is to serve rogue ads from compromised business accounts on the internet.

Fortunately, Meta has detected and stopped these threats, including previously unreported malware families. Still, the company’s security team has noticed rapid adversarial adaptation in response to their detection, which means it’s more important than ever to stay alert to these threats.

Malicious browser extensions

Meta has identified about ten malware families attacking people through the use of ChatGPT and other similar themes. As if that weren’t enough, threat actors are creating malicious browser extensions that claim to offer ChatGPT tools to trick people into downloading malware. These malicious extensions are promoted both on social media and in sponsored search results. Some of the extensions even include ChatGPT along with the malware to avoid suspicion from official web stores.

It is important to take proper precautions when browsing the web. Be wary of:

  • browser extensions that claim to offer ChatGPT tools;
  • sponsored ads on social media and in search results;
  • any link, website or app that looks suspicious.

Ducktail and NodeStealer: how the attacks work

Ducktail: more than 1,000 malicious URLs blocked

Meta announced that it has blocked more than 1,000 malicious ChatGPT-related URLs on its platforms, while sharing these URLs with industry partners. This effort was necessary due to the Ducktail operation, which has been targeting Facebook users since 2021. This operation spoofs ChatGPT to steal browser cookies, hijack logged-in Facebook sessions, and access sensitive information from its victims’ accounts, including two-factor authentication code and location data.

NodeStealer: cookie theft and account hijacking

In January of this year, Meta researchers discovered the NodeStealer malware, which steals cookies to break into Facebook, Gmail and Outlook accounts. Quickly, the company was able to identify the malware and take steps to stop its action, helping potential victims recover their accounts. To do so, removal requests were sent to third-party registrars, hosting providers, and applications used by threat actors.

Since February 27 of this year, Meta researchers have not observed new malware samples from the NodeStealer family. However, the company continues to monitor any potential future activities, including the growing threat of generative artificial intelligence.

BlackFog’s warning and how to protect data

Security experts at the company BlackFog have issued a warning about the threat of ChatGPT, highlighting its potential to produce malicious code. The company is closely monitoring how artificial intelligence can be used as bait on social media. According to BlackFog, ChatGPT can be used to exfiltrate data and create phishing websites, with the purpose of stealing credentials and installing malware on devices. As a result, cybersecurity measures will need to keep pace with this emerging technology.

Traditional approaches such as EDR-based defense and antivirus have proven ineffective against these modern ransomware variants. To protect data, it is necessary to use newer technologies to prevent data exfiltration in the first place. Without the ability to exfiltrate the information, attackers have no way to extort victims and have nothing to gain. In addition, the due diligence of users is essential to avoid falling for spoofing campaigns.

Alexandre Morelli
Head of Cybersecurity LATAM at Stefanini

LET'S TALK

From reflection to practice

How do these questions connect to your business right now? Talk to Stefanini about strategy, technology and customer experience.

Talk to Stefanini