Malicious Google search ads such as OpenAI ChatGPT and Midjourney are used to direct users to frame websites as part of a BATLOADER campaign designed to deliver the Redline Stealer malware.
“Both AI services are very popular but lack first-party standalone apps (ie, Midjourney uses Discord while users interface with ChatGPT via the web interface)” eSentire he said. By analyzing.
“This vulnerability has been exploited by threat actors seeking to impersonate AI app searchers as websites promoting fake apps.”
BATLOADER is a download loader malware that spreads when viewed by users searching for certain keywords on search engines, and when clicked, redirects to fraudulent landing pages that host the malware.
The installer file, by eSentire, is manipulated by downloading and installing RedLine Stealer from a remote server (ChatGPT.exe or midjourney.exe) and a PowerShell script (Chat.ps1 or Chat-Ready.ps1).
Once the installation is complete, the binary will use Microsoft Edge Web View2 To install chat.openai[.]com or www.midjourney[.]com – the legitimate chatgpt and midjourney urls – in the popup so that no red flags are raised.
The adversary used chatgpty and midjourney-centric lures to deliver malicious ads and eventually drop the RedLine Stealer malware, which was highlighted by Trend Micro last week.
This isn’t the first time the operators behind BATLOADER have capitalized on the AI craze to spread malware. In the year In March 2023, Esentire detailed a similar set of attacks that led to the deployment of ChatGPT decoys. Vidar stole and Ursnif.
The cyber security company has indicated that the abuse of Google search ads has dropped since the beginning of 2023, indicating that the tech giant is taking active steps to curb the exploit.
Defense against insider threats: Master SaaS security posture management
Worried about insider threats? We’ve got you covered! Join this webinar to explore practical strategies and security secrets for SaaS security deployment management.
The development coincides with widespread phishing and spoofing campaigns, in which actors are trying to make money by using these AI tools to distribute malware and other fake applications.
In a similar study by security provider Sophos, a collection of FleeceGPT-related fleeceware applications in Google Play and the Apple App Store – collectively dubbed FleeceGPT – forced users to sign up for unwanted registrations.
Sophos researchers Jagadesh Chandraya and Sean Gallagher said: “Fluffware apps stay on the edge of Apple and Google’s terms of service and are less likely to be rejected during review and allowed into the app store because they don’t access personal data or try to circumvent the platform’s security.” he said..
In recent weeks, checkpointmeta and Palo Alto Networks Division 42 they have warned Adding fraud functionality by impersonating the ChatGPT service to collect users’ credit card details, commit credit card fraud, and steal victims’ Facebook account details with a copycat bot web browser extension.
From November 2022 to early April 2023, Section 42 reported a 910% increase in monthly registrations for domains associated with ChatGPT.
Essentir, in January, traced the identities of key malware-as-a-service (MaaS) operators to an individual in Montreal, Canada. The second threat actor associated with the group was from that time Identified As a Romanian citizen who goes by the nickname Jack.