A sinister development in the world of artificial intelligence has emerged. Meet WormGPT, a chatbot specifically designed to assist cybercriminals in their nefarious activities. The developer of this malicious AI, who remains anonymous, is now selling access to the program on a well-known hacking forum, as reported by email security provider, SlashNext.
Unlike its benign counterpart, ChatGPT, or even Google's Bard, WormGPT lacks any ethical constraints. This makes it an ideal tool for hackers seeking to create malware and orchestrate phishing attacks with relative ease. The hacker initially introduced WormGPT back in March but officially launched it just last month.
In a blog post, SlashNext expressed its concern over the emergence of custom modules like WormGPT, tailored explicitly for malicious purposes. The ease of use and potential for illegal activities provided by this chatbot are alarming, as it empowers anyone to engage in cybercrime from the comfort of their own home.
To demonstrate the capabilities of WormGPT, its developer uploaded screenshots showcasing how the bot can produce malware written in the Python coding language. Furthermore, it can provide tips and guidance on crafting sophisticated and convincing malicious attacks.
The development of WormGPT involved training an older open-source language model called GPT-J from 2021 on data related to malware creation, resulting in the birth of this malicious chatbot.
In a test conducted by SlashNext, WormGPT was put to the challenge of crafting a convincing business email compromise (BEC) scheme — an infamous type of phishing attack. The results were nothing short of unsettling. The chatbot produced an email that was both persuasive and strategically cunning, showcasing its potential to carry out sophisticated phishing and BEC attacks. The language used in the message was professional and devoid of any spelling or grammar errors, which are typical red flags for phishing emails.
SlashNext expressed deep concern over WormGPT's potential threat to cybersecurity. The absence of ethical boundaries and the potency of generative AI technologies in the wrong hands, even those of novice cybercriminals, underscore the alarming risks posed by such AI developments.
Despite the ominous capabilities of WormGPT, its access doesn't come cheap. The developer offers access to this malicious tool for 60 Euros per month or 550 Euros per year. There has been a complaint from one buyer, criticizing the bot's performance, but the overall existence of WormGPT serves as a grave reminder of how generative AI programs could fuel cybercrime, especially as they continue to mature and evolve.
In conclusion, WormGPT's emergence serves as a stark warning of the dark potential of artificial intelligence when wielded for malicious purposes. As the cybersecurity landscape evolves, it becomes ever more crucial for researchers, developers, and the industry to stay vigilant and adopt robust measures to mitigate the threats posed by such malevolent AI advancements.
