Outsmarting FraudGPT - Personal Safeguards Explored

The rise of artificial intelligence (AI) is a double-edged sword. Like every transformative technology, AI offers immense potential but also enormous risks. Despite the emphatic push to regulate AI, threat actors seem to have gotten ahead of the curve.

MUO VIDEO OF THE DAY

SCROLL TO CONTINUE WITH CONTENT

A new ChatGPT-styled tool, FraudGPT, is gaining traction among cybercriminals, allowing them to automate and better execute a large part of their fraud operations. Anyone can become a victim, so it is important to stay informed. Here’s everything we know about FraudGPT so far.

Disclaimer: This post includes affiliate links

If you click on a link and make a purchase, I may receive a commission at no extra cost to you.

What Is FraudGPT?

FraudGPT is an AI tool powered by a large language model that is particularly fine-tuned to help cyber criminals commit cybercrime. The subscription-based AI tool allows threat actors to facilitate their criminal activities like carding, phishing, and malware creation.

Although details about the tool remain very limited, researchers from Netenrich, a security research firm, have uncovered several ads on the dark web advertising the tool. According to the research report from Netenrich , the subscription fees range between $200 per month to $1,700 yearly.

To better picture the tool, you can think of FraudGPT as ChatGPT but for fraud. But how exactly does FraudGPT work, and how are cybercriminals using it?

How Does FraudGPT Work?

Image Credit: Freepik

At its core, FraudGPT is not significantly different from any tool that is powered by a large language model. In other words, the tool itself is an interactive interface for criminals to access a special kind of language model that has been tuned for committing cyber crimes.

Still don’t get it? Here’s a rough idea of what we are talking about. In the early days of ChatGPT’s launch, the AI chatbot could be used to do just about anything, including helping cyber criminals create malware. This was possible because ChatGPT’s underlying language model was trained on a dataset that likely contained samples of a wide range of data, including data that could help a criminal venture.

Large language models are typically fed everything from the good stuff, like science theories, health information, and politics, to the not-so-good ones, like samples of malware code, messages from carding and phishing campaigns, and other criminal materials. When dealing with the kind of datasets needed to train language models like the kind that powers ChatGPT, it is almost inevitable that the model would contain samples of unwanted materials.

Despite typically meticulous efforts to eliminate the unwanted materials in the dataset, some slip through, and they are usually still enough to give the model the ability to generate materials to facilitate cybercrime. This is why with the right prompt engineering , you can get tools like ChatGPT, Google Bard, and Bing AI to help you write scam emails or computer malware.

If tools like ChatGPT can help cyber criminals commit crimes despite all efforts to make these AI chatbots safe, now think of the power a tool like FraudGPT could bring, considering it was specifically fine-tuned on malicious materials to make it suitable for cybercrime. It’s like the evil twin of ChatGPT put on steroids.

So, to use the tool, criminals could just prompt the chatbot as they’d do with ChatGPT. They could ask it to, say, write a phishing email for Jane Doe, who works at company ABC, or maybe ask it to write malware using C++ to steal all the PDF files from a Windows 10 computer. Criminals would basically just come up with evil mechanization and let the chatbot do the heavy lifting.

How Can You Protect Yourself From FraudGPT?

Despite being a new kind of tool, the threat posed by FraudGPT is not fundamentally different. You could say it introduces more automation and efficiency to already established methods of executing cybercrime.

Criminals using the tool would, at least theoretically, be able to write more convincing phishing emails, better plan scams, and create more effective malware, but they’d mostly still rely on the established ways of executing their nefarious plans. As a result, the established ways to protect yourself still apply:

• Be wary of unsolicited messages asking for information or directing you to click on links. Do not provide information or click links in these messages unless you verify the source.
• Contact companies directly using an independent, verified number to check legitimacy. Do not use contact info provided in suspicious messages.
• Use strong, unique passwords that are hard to crack and embrace two-factor authentication when available for all accounts. Never share passwords or codes sent to you.
• Regularly check account statements for suspicious activity.
• Keep software updated and use antivirus or anti-malware tools.
• Shred documents containing personally identifying or financial information when no longer needed.

For more on how to protect yourself, read our guide on how to protect yourself in the era of AI .

Stay Informed to Protect Yourself

The emergence of tools like FraudGPT reminds us that despite all the good that AI can do for us, it still represents a very potent tool in the hands of threat actors.

As governments and large AI firms go on a frantic race to enact better ways to regulate AI, it is important to be aware of the threat that AI currently poses and take the necessary precautions to protect yourself.

SCROLL TO CONTINUE WITH CONTENT

A new ChatGPT-styled tool, FraudGPT, is gaining traction among cybercriminals, allowing them to automate and better execute a large part of their fraud operations. Anyone can become a victim, so it is important to stay informed. Here’s everything we know about FraudGPT so far.

Also read: