FakeBat Impersonates Midjourney, ChatGPT in Drive-by Cyberattacks

Adversaries don’t work 9-5 and neither do we. At eSentire, our 24/7 SOCs are staffed with Elite Threat Hunters and Cyber Analysts who hunt, investigate, contain and respond to threats within minutes.

We have discovered some of the most dangerous threats and nation state attacks in our space – including the Kaseya MSP breach and the more_eggs malware.

Our Security Operations Centers are supported with Threat Intelligence, Tactical Threat Response and Advanced Threat Analytics driven by our Threat Response Unit – the TRU team.

In TRU Positives, eSentire’s Threat Response Unit (TRU) provides a summary of a recent threat investigation. We outline how we responded to the confirmed threat and what recommendations we have going forward.

Here’s the latest from our TRU Team…

This blog was updated in September 2023 based on research from TRU that identifies this as FakeBat. This blog originally identified this threat as BatLoader.

What did we find?

In early May, eSentire Threat Response Unit (TRU) identified an ongoing FakeBat campaign using Google Search Ads to deliver imposter web pages for ChatGPT and Midjourney:

• ChatGPT is an Artificial Intelligence chatbot released in November 2022 and has since exploded in popularity.
• Midjourney is a generative AI service in which users can submit text prompts to generate images.

Both AI services are extremely popular but lack first-party standalone apps (i.e., users interface with ChatGPT via their web interface while Midjourney uses Discord).

This vacuum has been exploited by threat actors looking to drive AI app-seekers to imposter web pages promoting fake apps.

In its latest campaign, FakeBat is using MSIX Windows App Installer files to infect devices with Redline Stealer. This is not the first time FakeBat has targeted users searching for AI tools. In February 2023, TRU identified a series of newly registered FakeBat domains, which included chatgpt-t[.]com.

Overview of Redline Infection from Imposter Ad for ChatGPT

Initial Download

In this case, the infection was traced to a Google search for “chatbpt”, leading to an imposter download page for ChatGPT hosted on hxxps://pcmartusa[.]com/gpt/:

The download link led to advert-job[.]ru then job-lionserver[.]site for the final payload.

Chat-GPT-x64.msix

Chat-GPT-x64.msix (md5hash: 86a9728fd66d70f0ce8ef945726c2b77) is a Windows App Package format used for installing applications.

Windows requires all files that make up an MSIX app to be signed with a common signature. This package is digitally signed by ASHANA GLOBAL LTD:

Examining the contents of the package, we see the various assets used during installation:

Looking at the AppXManifest file, we can see the package was created by a Russian speaker using Advanced Installer version 20.2 with a professional license:

Opening the package in AdvancedInstaller, we can see the application will launch both an executable (ChatGPT.exe) and a PowerShell script (Chat.ps1).

The installer will also use the ChatGPT logo and target Windows desktop versions between October 2018 Update – 1809 and October 2022 Update – 22H2.

Clicking the installer file launches the Windows App Installer wizard:

The file signature is crucial for MSIX packages, the installer will not let you proceed without a signature from a trusted certificate (Windows 10 requires all applications to be signed with a valid code signing certificate.)

During installation, Chat.ps1 and ChatGPT.exe are executed under the context of aistubx64.exe.

Chat.ps1 is a basic PowerShell download cradle. In this case it downloads and loads Redline Stealer from adv-pardorudy[.]ru into memory as an assembly. The script also performs two requests to the C2 panel:

• Start.php: records the start time of the infection along with the victim’s IP address.
• Install.php: records successful installation of the payload at adv-pardorudy[.]ru along with the time and IP address of the victim.

These actions are taken for statistical tracking purposes by operators, allowing them to easily identify successful victims and track metrics around specific campaigns or themes.

This Redline sample (md5hash 7716F2344BCEBD4B040077FC00FDB543) is configured to connect to IP 185.161.248[.]81 using the Bot ID “ChatGPT_Mid”, a reference to the two lures used in this campaign (ChatGPT and MidJourney).

Examining ChatGPT.exe, TRU observed that the executable uses Microsoft Edge WebView2 to load https://chat.openai.com/ in pop-up window post-installation.

The primary functionality is to redirect the user’s attention and assure them that they have installed a valid application. The result is a popup window containing the real ChatGPT web page embedded in a barebones browser window. Other functionality of this executable is not
immediately clear.

Redline Infection from Imposter Ad for Midjourney

In a separate case in May 2023, TRU observed a similar infection scheme promoting an imposter page for Midjourney. This resulted in the user downloading Midjourney-x64.msix, a Windows Application Package also signed by ASHANA GLOBAL LTD.

In this case, the installer executes an obfuscated PowerShell script (Chat-Ready.ps1), which ultimately is identical to the script shown in Figure 13 albeit with a different C2 domain.

What did we do?

• TRU conducted proactive threat hunts across our customers’ environments globally for the identified app packages.
• We deployed new detection content to identify MSIX app package abuse.
• Our team of 24/7 SOC Cyber Analysts alerted the impacted customers and provided remediation guidance and support.

What can you learn from this TRU Positive?

• Generative AI technologies and chatbots have exploded in popularity in 2023. Unfortunately, as system administrators seek ways to control access to these platforms, users may seek out alternative ways to gain access.
• Threat actors have been keen to exploit the popularity of these tools, promising unrestrictive access.
• Our telemetry shows Google Search Ads abuse (explained here) peaked in popularity in Q4 2022 and early 2023. The success rate has diminished, suggesting Google has tamped down on abuse of their ad service. However, this recent campaign shows malicious ads can still slip by moderators and deliver victims malware.
• This campaign bears several similarities to previously identified FakeBat activity:
  • Impersonating major brands and services using Google Search Ads.
  • Use of AdvancedInstaller for creating installation packages.
  • Payload site job-lionserver[.]site previously attributed to FakeBat.
  • Information stealing malware payloads.

Recommendations from our Threat Response Unit (TRU) Team:

Indicators of Compromise

eSentire’s Threat Response Unit (TRU) is a world-class team of threat researchers who develop new detections enriched by original threat intelligence and leverage new machine learning models that correlate multi-signal data and automate rapid response to advanced threats.

If you are not currently engaged with an MDR provider, eSentire MDR can help you reclaim the advantage and put your business ahead of disruption.

Learn what it means to have an elite team of Threat Hunters and Researchers that works for you. Connect with an eSentire Security Specialist.

Source: https://www.esentire.com/blog/batloader-impersonates-midjourney-chatgpt-in-drive-by-cyberattacks