Research shows that ChatGPT can be used to create malicious code

The biggest problem with such AI code generators is that natural language processing (NLP) engines can lower the barrier to entry for malicious hackers. With code generators that don’t require the user to be proficient in coding, any user can collate the logical flow of information used in a malicious tool from the open web and use the same logic to generate syntax for malicious tools.

Presenting the issue, Check Point introduces how the AI ​​code generator is used to prototype the basic code for the phishing email scam, and applies follow-up instructions in plain English to further improve the code. In what the attackers have demonstrated, therefore, any user with malicious intent can create an entire hacking campaign using these tools.

Sergey Shykevich, threat intelligence team manager at Check Point, says that tools like ChatGPT have “the potential to dramatically change the cyber threat landscape”.

“Hackers can also duplicate malicious code using ChatGPT and Codex. AI technologies represent another step in the dangerous evolution of increasingly sophisticated and effective cyber capabilities,” he added.

To be sure, while open source language models can also be used to create cyber defense tools, the lack of protection in terms of usage to create malicious tools can be alarming. motion. Check Point notes that while ChatGPT has stated that using its platform to create hacking tools is “against” its policy, there are no restrictions preventing it from doing so.

This isn’t the first time an AI language and image rendering service has shown potential for abuse. Lensa, an AI-based image editing and retouching tool by US-based Prisma, also highlighted the lack of filters based on body and nudity that can lead to offending images. privacy is created by an individual without consent.