How WormGPT, An Evil Rival To ChatGPT, Is Used By Hackers To Plant Exceptional Email Attacks | Explained

JE Technology Desk: We all have heard of ChaGPT for its helpful AI capabilities, but do you know a tool called WormGPT exists in the wild that is actively employed by cybercriminals to launch sophisticated Business Email Compromise (BEC) attacks? Slashnext, in a recent blog post, highlighted real-life instances from several cybercrime platforms that explored the modus operandi of such attacks and how generative AI is acting as a magic trick to fool people in unique and near-perfect ways.

WormGPT: A Hacker's Best Friend

Designed to execute malicious activities, WormGPT is tagged a 'blackhat' rival to the GPT alternatives. It draws power from GPTJ, a large language model from 2021. GPTJ sports features such as context memory, code editing and infinite character limit and was trained through malicious data. As a result, the model can produce genuine-looking emails that can mesmerise the recipient at first glance. This ChatGPT clone with no ethical limits reminds us of the downsides of AI. 

Also Read: Evil Chatbot ChaosGPT Was Asked To 'Destroy Humanity'; Here's How It Responded

How do these sophisticated AI attacks work?

Image: Pexels

As AI tools get smarter these days, they are being used to generate texts that sound natural and are free from errors. Cyber fraudsters deploy these tools to prepare fake emails and customise them to hypnotise the receiver into believing them as genuine.

The report shared the conversation wherein a cyber fraudster presented an example of how AI is helping them refine phishing attacks targeting a business. This scammer recommends framing the email in the native language and translating it. Finally, ChatGPT is used to sprinkle the formal touch. This implies that the fraudster's proficiency in a language no longer matters with the arrival of AI.

Also Read: AI Appreciation Day: From AI Historian To ChatGPT, Check Impressive Tools Announced Last Year

In another case, cybercriminals offer what can be termed 'jailbreaks'. These carefully crafted prompts trick ChatGPT to surrender sensitive details, generating harmful codes and creating inappropriate content. These bad actors also use custom modules like ChatGPT for other scandalous purposes. These modules are passed further to other people involved in the trade.

How Generative AI is helping hackers in BEC Attacks?

Image: Pexels

The emerging technology has conferred hackers who are less fluent in technology and languages with the power to use unbeatable grammar in Email attacks.

To understand the working, the platform entered a prompt to frame a BEC attack addressing an account manager to "urgently pay an invoice" that should mimic a business CEO as a sender. Take a look for yourself at the capabilities of WormGPT.

Image: Slashnext