Fake ChatGPT Custom GPTs Used to Spread Remote Access Trojans via ClickFix Lures
Written by AI from the linked story. Not reviewed by an editor. How this works
Threat actors created malicious ChatGPT Custom GPTs that impersonate legitimate tools, tricking users into visiting attacker-controlled sites that deploy ClickFix social engineering lures. Victims are manipulated into executing a complex MSI installer chain that leverages DLL sideloading and shellcode execution to install a Remote Access Trojan (RAT). This attack demonstrates how emerging AI platforms can be weaponized as a distribution vector when platform-level controls and user vigilance are insufficient. With at least 40 confirmed infections, the campaign highlights the real-world impact of combining AI-platform abuse with layered malware delivery techniques. Organizations must treat AI tool usage as an expanded attack surface requiring both technical controls and user education.
Tactical Insight
Immediate actions
- Block or restrict end-user access to unvetted ChatGPT Custom GPTs through organizational AI usage policies.
- Deploy endpoint detection and response (EDR) tools capable of detecting DLL sideloading, suspicious MSI execution, and shellcode injection behaviors.
User awareness & training
- Train employees to recognize ClickFix and social engineering lures that prompt them to run commands or install software from browser prompts.
- Establish a clear reporting procedure so users can quickly flag suspicious AI-generated content or unexpected software installation requests.
Long-term improvements
- Implement application whitelisting to prevent unauthorized MSI installers and unsigned DLLs from executing on endpoints.
- Develop and enforce an AI tool governance policy that vets and approves sanctioned AI platforms before organizational use.
- Continuously monitor endpoint telemetry and network traffic for RAT command-and-control (C2) communication patterns using SIEM correlation rules.