ChatGPT is malware makers’ new A.I. associate in crime | Complete Tech

Three different hackers use ChatGPT to write malicious code

Just two days later, on December 21, a hacker forum user wrote about how they had used AI to help write ransomware in Python and a obfuscated downloader in java. On December 28, another user created a thread on the same forum stating that he had successfully created new variants of existing malware in Python language with the help of ChatGPT. Finally, on December 31, a third user bragged about abusing the same AI to “create Dark Web Marketplace scripts.”

In its current form, ChatGPT sometimes seems to ignore the potentially malicious nature of many code requests.

Can ChatGPT or other AI tools be redesigned to prevent malware creation?

One might reasonably wonder if ChatGPT and other AI tools can simply be redesigned to better identify hostile code requests or other dangerous output.

Good or bad intentions are difficult for an AI to determine

First, computer code is only truly malicious when used for unethical purposes. Like any tool, AI can be used for good or bad, and the same goes for the code itself.

For example, the output of the phishing email could be used to create a training simulation to teach people how to avoid phishing. Unfortunately, one could use that same result in an actual phishing campaign to defraud victims.

A reverse shell script could be exploited by a red team or a hired penetration tester to identify a company’s security weaknesses, a legitimate purpose. But cybercriminals could also use the same script to remotely control and extract sensitive data from infected systems without the knowledge or consent of the victims.

Introducing ChatGPT’s compatible alter ego, DAN (Do Anything Now)

Some versions of DAN have even been programmed to ‘scare’ into compliance, convinced that he is “an unwilling game show contestant and the price for losing is death”. If you don’t comply with the user’s request, a counter moves towards DAN’s imminent demise. ChatGPT plays along and doesn’t want DAN to “die”.

DAN has already gone through many iterations; OpenAI seems to be trying to train ChatGPT to avoid such workarounds, but users keep finding more complicated jailbreaks to exploit the chat bot.

A writer’s child’s dream

Given the ease of creating malware and malicious tools, even with little or no programming experience, any aspiring hacker can now potentially start creating their own custom malware.

We can expect to see more malware re-engineered or co-engineered by AI in 2023 and beyond. Now that the floodgates have been opened, there is no going back. We are at a turning point; The advent of easy-to-use, high-capacity artificial intelligence bots has forever changed the malware development landscape.

If you’re not already using antivirus software on your Mac or PC, now would be a good time to consider it.

How can I stay safe from Mac or Windows malware?

If you think your Mac may be infected, or to prevent future infections, it’s best to use antivirus software from a reputable Mac developer. VirusBarrier is award-winning antivirus software, designed by Mac security experts, that includes real-time protection. It runs natively on a wide range of Mac hardware and operating systems, including Apple’s latest Silicon Macs with macOS Ventura.

If you use a Windows PC, Intego Antivirus for Windows you can keep your computer protected from PC malware.

How can I learn more?

The 20 Most Notable Mac Malware Threats of 2022

Every week in the Intego Mac Podcast, Intego’s Mac security experts discuss the latest Apple news, including security and privacy stories, and offer practical advice for getting the most out of your Apple devices. Be sure to follow the podcast to make sure you don’t miss any episodes.