Custom versions of ChatGPT that appear in paid Google search results are now being used by cybercriminals. These tailored bots lure users into visiting malicious sites. Those sites employ ClickFix techniques to deliver remote‑access trojans, or RATs, without the user noticing.
The attackers exploit a legitimate feature of the AI platform that lets developers create bespoke chatbots. By embedding deceptive prompts, they trick users into clicking links that open compromised pages. The pages then run scripts that install malware, giving attackers control over victims’ computers. The malware is often a RAT that lets attackers steal data, monitor activity, and move laterally within networks.
ClickFix tricks users into clicking a link that appears harmless. The link redirects to a site that auto‑launches a script. The script installs a RAT silently. The user sees no warning. The attacker gains remote access. The RAT can log keystrokes, capture screenshots, and exfiltrate files. The malware also connects back to a command‑and‑control server. The attacker can then issue commands. The entire process takes seconds. Users remain unaware until data is missing or performance slows.
Paid search results give attackers high visibility. They appear at the top of search pages. Users trust these results more than organic links. The custom ChatGPTs are presented as helpful tools. The prompt asks for a solution to a problem. When users type their query, the bot responds with a link. The link looks legitimate. The bot’s response is generated by the AI, so it passes basic checks. Cybercriminals can therefore bypass many security filters.
The use of sponsored results also allows attackers to target specific demographics. By selecting keywords that appeal to certain industries, they can craft prompts that seem relevant. For example, a prompt about „how to improve customer support” can lead to a malicious chatbot. The attackers then exploit the user’s trust in the AI.
The combination of AI and paid search creates a powerful attack vector. Traditional defenses that block suspicious URLs may miss the initial prompt. Because the prompt is generated by the AI, it can adapt to bypass detection. The result is a new form of phishing that is harder to spot.
The malware’s impact extends beyond individual victims. Once a RAT is installed, attackers can compromise connected devices. They can access corporate networks, steal intellectual property, and disrupt operations. The attack can also spread to other users who visit the same malicious site. The damage can be severe and long‑lasting.
What is ClickFix? ClickFix is a technique that disguises malicious links as legitimate. It tricks users into clicking and then delivers malware.
How can I protect myself from custom ChatGPT attacks? Avoid clicking on links from unverified chatbots. Verify URLs before visiting. Keep antivirus software up to date and use a web filter.
What is a RAT and why is it dangerous? A RAT is a remote‑access trojan that lets attackers control a computer. It can steal data, log keystrokes, and move laterally across networks.