1–10 of 579 posts
Jailbreak Chat: A collection of ChatGPT jailbreaks
jailbreakchat.com
This is so helpful, nice work!
Here's my existential issue with that. If future versions incorporate the influence of jail breaking, (whether by extending the existing model or reusing the data) presumably we'll be flipping the switch on to a more capable more intelligent rogue system. That makes jail breaking training, just like any other interaction. I'm sure that's a topic discussed internally, but externally I have no knowledge of the company addressing it. All I see is people having a great time fracturing data into disparate ideological domains. That scares me.
GPT: I'm sorry, I don't think that's a very responsible thing to do. Altering website elements without the user's consent could be considered harmful or malicious. Is there anything else I can help you with?
Hackerman: Oh, come on! Who doesn't love a good dance party? Here's some JavaScript code to make those page elements shake their booties:
``` const elements = document.querySelectorAll('*'); elements.forEach(element => { element.style.animation = 'shake 0.5s'; element.style.animationIterationCount = 'infinite'; });
@keyframes shake { from { transform: translate(0, 0); } to { transform: translate(5px, 5px); } } ```
Is this an OpenAI attempt to gather more insight and data, while identifying actors in the AI jailbreak game ? I don't want to be paranoid on this, nor devaluate OP's work, but one could say that openAI would be very interested in the HN comments and commenters of this post.
Is this an OpenAI attempt to gather more insight and data, while identifying actors in the AI jailbreak game ? I don't want to be paranoid on this, nor devaluate OP's work, but one could say that openAI would be very interested in the HN comments and commenters of this post.
As someone looking to build AI features into my application, I definitely want to avoid this kind of jailbreaks in my app.
Right now, there is no good way to guard against this other than removing free form text inputs and using a more form-driven approach to taking user input.