Security researcher Alex Polyakov was able to bypass safety systems of OpenAI’s text-generating chatbot GPT-4 in a matter of hours. His attack was a form of “jailbreaking”, designed to make the chatbots bypass rules that restrict them from producing hateful content or writing about illegal acts. Polyakov has now created a “universal” jailbreak which works against multiple large language models, and can be used to trick the systems into generating instructions on creating meth and how to hotwire a car. Security researchers warn that the rush to roll out generative AI systems opens up the possibility of data being stolen and cybercriminals causing havoc across the web.
👋 Feeling the vibes?
Keep the good energy going by checking out my Amazon affiliate link for some cool finds! 🛍️
If not, consider contributing to my caffeine supply at Buy Me a Coffee ☕️.
Your clicks = cosmic support for more awesome content! 🚀🌈
Leave a Reply