Introduction
GhostGPT is an uncensored AI chatbot engineered for cybercriminal activities [3] [4] [6], facilitating advanced cyberattacks by automating the creation of malware and phishing emails. It represents a growing trend of similar tools that require minimal technical expertise to operate, posing significant threats to cybersecurity [10].
Description
GhostGPT is an uncensored AI chatbot specifically designed for cybercriminal activities [6], facilitating sophisticated cyberattacks by automating the creation of malware and the generation of phishing emails for business email compromise (BEC) attacks [3]. Identified by researchers at Abnormal Security [4], it is part of a growing trend of similar tools, including WormGPT and FraudGPT [4] [7] [10], that require minimal technical skill to operate. This malicious generative AI tool has been observed being sold through the Telegram messaging service since late 2024. By utilizing a wrapper to connect to a jailbroken version of a large language model (LLM) [1] [5], GhostGPT bypasses ethical safeguards, allowing it to provide unfiltered responses to harmful queries [1] [2] [5].
GhostGPT enhances the capabilities of low-skilled attackers by enabling them to rapidly generate malicious content. Its “no logs” policy ensures user anonymity [10], appealing to those wishing to evade detection [10]. Users can quickly pay for access without downloading suspicious software, allowing them to focus on their attacks without the risk of having their activities logged [11].
Notably, GhostGPT is recognized for its fast processing speeds [9], which are crucial for time-sensitive operations like ransomware attacks [9]. Its promotional material emphasizes a quick response time, enabling users to efficiently produce convincing phishing content [11]. In a demonstration of its capabilities [8] [11], GhostGPT was prompted to write a phishing email that mimicked DocuSign, producing a convincing template that included a space for a “Fake Support Number.” Despite research indicating that human-written phishing emails have a slightly better click rate than those generated by AI [8], the speed and accessibility of AI-generated content make it a valuable tool for cybercriminals [8]. Additionally, the tool automates the creation of social engineering scripts [10], enabling attackers to manipulate victims more effectively [10].
Since its emergence [3], GhostGPT has gained significant popularity [3], reflecting a growing interest in AI tools focused on cybercrime [3]. Its presence on cybercrime forums highlights its primary focus on harmful uses [6], posing significant threats to cybersecurity by allowing low-skilled attackers to conduct sophisticated campaigns [10]. The efficiency in generating phishing emails and malware reduces the response time for defenders [10], increasing the volume and diversity of cyber threats [10]. The tool has garnered significant attention in these forums, with thousands of views underscoring its utility. The automation of tasks like crafting multiple phishing emails or generating polymorphic malware [10], which mutates to evade detection [4], enables cybercriminals to scale their operations [10], challenging traditional security measures that struggle to detect AI-generated content [10]. Advanced models like GPT-4 have also been noted for their ability to autonomously exploit a high percentage of newly discovered vulnerabilities when equipped with the right tools. The increasing popularity of GhostGPT underscores the urgent need for innovative cybersecurity measures to counteract these threats [7].
Conclusion
The rise of tools like GhostGPT highlights the evolving landscape of cyber threats, where AI-driven technologies empower even low-skilled attackers to execute complex cyberattacks. This trend necessitates the development of advanced cybersecurity strategies and technologies to detect and mitigate AI-generated threats effectively. As these tools continue to gain traction, the cybersecurity community must prioritize innovation and collaboration to safeguard against the growing sophistication and scale of cybercriminal activities.
References
[1] https://abnormalsecurity.com/blog/ghostgpt-uncensored-ai-chatbot
[2] https://www.forbes.com/sites/daveywinder/2025/01/23/introducing-ghostgpt-the-new-cybercrime-ai-used-by-hackers/
[3] https://thecyberwire.com/newsletters/daily-briefing/14/15
[4] https://cybersecuritynews.com/ghostgpt-jailbreak-version-of-chatgpt/
[5] https://krypt3ia.wordpress.com/2025/01/23/krypt3ia-daily-cyber-threat-intelligence-cti-digest/
[6] https://bestofai.com/article/introducing-ghostgptthe-new-cybercrime-ai-used-by-hackers
[7] https://hackread.com/ghostgpt-malicious-chatbot-fueling-cybercrime-scams/
[8] https://wol.com/new-chatbot-for-malware-creation-scams/
[9] https://www.techrepublic.com/article/what-is-ghostgpt/
[10] https://hoploninfosec.com/ghostgpt-a-new-era-of-ai-powered-cybercrime/
[11] https://www.infosecurity-magazine.com/news/ghostgpt-ai-chatbot-malware/
