Cybersecurity researchers were able to bypass security features on ChatGPT by roleplaying with it, getting the bot to write password-stealing malware.

Cybersecurity researchers were able to bypass security features on ChatGPT by roleplaying with it, getting the bot to write password-stealing malware.

Hackers have infiltrated a tool your software development teams may be using to write code. Not a comfortable place to be.

Bad actors, unconfined by ethical boundaries, recently released two large language models designed to help fraudsters write phishing prompts and hackers write malware.

They can also ask it to write malware. In the hands of hackers, ChatGPT can be really good at creating hard-to-detect malware. If they combine it with machine learning models, it's even more powerful.

Very few malware writers are able to write such clean code that can install on a variety of hardware systems, assess their new environments and download the modules they need to successfully ...

* Students with experience writing malware would be unemployable by anti-virus firms, always concerned about the widespread rumor that they engage in writing viruses for profit.

Cybersecurity researchers were able to bypass security features on ChatGPT by roleplaying with it, getting the bot to write password-stealing malware.