Close Menu
Versa AI hub
  • AI Ethics
  • AI Legislation
  • Business
  • Cybersecurity
  • Media and Entertainment
  • Content Creation
  • Art Generation
  • Research
  • Tools

Subscribe to Updates

Subscribe to our newsletter and stay updated with the latest news and exclusive offers.

What's Hot

UK companies compete to embed AI into their enterprise workflows.

May 31, 2025

Deepseek’s latest AI model is a “big step back” for free speech

May 31, 2025

Mewtant Inc. announces the innovative Pixai Studio Ghibli AI Image Generator for free

May 30, 2025
Facebook X (Twitter) Instagram
Versa AI hubVersa AI hub
Saturday, May 31
Facebook X (Twitter) Instagram
Login
  • AI Ethics
  • AI Legislation
  • Business
  • Cybersecurity
  • Media and Entertainment
  • Content Creation
  • Art Generation
  • Research
  • Tools
Versa AI hub
Home»Media and Entertainment»Mandiant finds over 30 fake AI websites spreading malware
Media and Entertainment

Mandiant finds over 30 fake AI websites spreading malware

versatileaiBy versatileaiMay 27, 2025No Comments3 Mins Read
Share Facebook Twitter Pinterest LinkedIn Tumblr Reddit Telegram Email
#image_title
Share
Facebook Twitter LinkedIn Pinterest Email
According to Mandiant, fake AI video generation websites promoted via Facebook and LinkedIn ads have been found to spread a variety of malware payloads with reconnaissance and backdoor features. The campaign has been ongoing since at least mid-2024, mimicking popular AI tools like Luma AI, Kinva Dream Lab, Canva Dream Lab, Luma AI like Canva Dream Lab, and Canva Dream Lab. The threat actor behind the campaign was tracked as UNC6032 by Mandiant, a part of Google Cloud, and is thought to be of Vietnamese origin. Millions of Facebook ads combine millions of views to point users towards malicious sites, with around 10 LinkedIn ads reported about 50,000 to 250,000 Imprene Thins, reporting an estimated 50,000 to 250,000 impressions. Morphisec, UNC6032’s fake website claims to provide free text-to-video image or video-to-video functionality, and ultimately provides a malicious executable file that is disguised as a .MP4 video file in response to user prompts or image uploads. “Mandiant has observed the various paths used by threat groups. Morphisec explains one of the other paths that this group uses.”

Malicious DLLs for Multi-Stage Attack Chain Side Road

In an analysis of one UNC6032 attack, in Mandiant’s threat defense, the last dropper tracked as starkveil was used to deploy three final Python-based payloads after the multi-stage attack chain. In this case, the file was found using some extensions to use some files of fake extensions to use Luma ai spoofing bour ai. The use of the name.Whitespace character, along with the default .mp4 Windows icon, tricks users into believing that the file is an AI-generated video they were expecting. You need to run Starkveil Dropper twice and you need to complete the attack chain. When you run the file, you will see an error window designed to allow the user to reopen the video. On the first run, Starkveil drops the embedded file into the C:\winsystem\ directory, and on the second run, Python Launcher’s Py.exe is invited to Python’s decodeaseaseedededeaseededeasededeasededeased as coilhatted.exe. Mandiant explained that the first stage Base85 encoded Python code decodes the second stage Python Bytecode. This second stage code will eventually run a legal digitally signed executable that will eventually be used to sideload the final launcher Heif.dll.

3 payload establishes backdoor persistence and sets stages of infosteel

The last three payloads are known as Grimpull, Xworm and Frostrift, the latter having both reconnaissance and backdoor functions, and Grimpull acts as a downloader to obtain subsequent payloads. Grimpull is surrounded by legitimate python.exe processes as Avcodec-61.dll and performs multiple checks to ensure that it is not running in a sandbox environment or a virtual machine (VM). The downloader then connects to the Command and Control (C2) server via TOR and periodically checks the .NET payload to decrypt and load it into memory. UNC6032 uses XWORM to collect system information, extend it to telegram chat, receives log keystrokes from an external C2 server, and commands from the ffplay.exe process that is kept in the FFPlay.exe process. This reconnaissance will help set the stage for future info-formula and subsequent attacks that may be facilitated by payloads acquired by Grimpull Downloader. Malware uses Xworm and Frostrift Autorun registry keys to establish backdoor persistence. and malware distribution. Check Point Research reported another similar campaign earlier this month. KlingAI spread the malware in PureHVNC rats. Added to AI video tools, the popular leading language model (LLMS) is also targeting impersonation, as seen in the Malvertising campaign reported by MalwareBytes in March, mimicking Deepseek.

author avatar
versatileai
See Full Bio
Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
Previous ArticleBipartisan group of state AGS reports on nix a decade-old moratorium on AI law
Next Article The UK deploys AI to increase Arctic security amid growing threats
versatileai

Related Posts

Media and Entertainment

SMPTE Bits by Bay is tackling the future impact of AI media

May 27, 2025
Media and Entertainment

Entertainment without boundaries: AI-Media and Lightning

May 27, 2025
Media and Entertainment

Most User Data Hungry AI Chatbots have been identified

May 27, 2025
Add A Comment
Leave A Reply Cancel Reply

Top Posts

The UAE announces bold AI-led plans to revolutionize the law

April 22, 20253 Views

The UAE will use artificial intelligence to develop new laws

April 22, 20253 Views

New report on national security risks from weakened AI safety frameworks

April 22, 20253 Views
Stay In Touch
  • YouTube
  • TikTok
  • Twitter
  • Instagram
  • Threads
Latest Reviews

Subscribe to Updates

Subscribe to our newsletter and stay updated with the latest news and exclusive offers.

Most Popular

The UAE announces bold AI-led plans to revolutionize the law

April 22, 20253 Views

The UAE will use artificial intelligence to develop new laws

April 22, 20253 Views

New report on national security risks from weakened AI safety frameworks

April 22, 20253 Views
Don't Miss

UK companies compete to embed AI into their enterprise workflows.

May 31, 2025

Deepseek’s latest AI model is a “big step back” for free speech

May 31, 2025

Mewtant Inc. announces the innovative Pixai Studio Ghibli AI Image Generator for free

May 30, 2025
Service Area
X (Twitter) Instagram YouTube TikTok Threads RSS
  • About Us
  • Contact Us
  • Privacy Policy
  • Terms and Conditions
  • Disclaimer
© 2025 Versa AI Hub. All Rights Reserved.

Type above and press Enter to search. Press Esc to cancel.

Sign In or Register

Welcome Back!

Login to your account below.

Lost password?