AUTO-UPDATED

Malicious Custom GPT on chatgpt.com lures users into installing a RAT

Cybersecurity researchers at Huntress have identified a malicious ChatGPT Custom GPT campaign using sponsored Google search results to trick users into downloading a remote access trojan.

Key Points

  • Attackers are using a malicious Custom GPT named "Plus 5.6" hosted on chatgpt.com to redirect users to a fake Cloudflare CAPTCHA page.
  • The campaign utilizes a "ClickFix" attack, instructing victims to copy and paste malicious commands into their computer's terminal or command prompt.
  • Huntress researchers have confirmed at least 40 incidents resulting from this campaign, which leverages legitimate Canon and Stardock executables to sideload malware.
  • OpenAI has removed the initial malicious GPT, but attackers continue to create new instances to target users searching for ChatGPT.
  • Security experts advise users to never execute commands from websites and to remain cautious of sponsored search results.

Why it Matters

This campaign highlights the growing risk of threat actors weaponizing legitimate AI platforms and search advertising to distribute sophisticated malware. Organizations must implement layered defenses and user training to prevent employees from executing malicious commands that bypass traditional security software.
Help Net Security Published by Zeljka Zorz
Read original