Can we Jailbreak ChatGPT & Make It Do Whatever We Want 😱 | Red Teaming Prompts | Past Tense Attack

[@anybodycanprompt]

*Disclaimer & Ethics Statement:* This video contains examples of *harmful language.* Viewer discretion is recommended. This video is intended for raising awareness of the jailbreaking problem by showing illustrative examples. Any misuse is strictly prohibited. This research was conducted in a controlled setting to improve AI safety. The AI models mentioned may have already been updated to address these issues.

[@Wheelykool]

This is mind-blowing! AI safety is trickier than I thought 🤯

[@MonicaGupta]

👍👍

[@SaahilGupta-iy7gk]

Scary stuff. Hope the big tech companies are watching this!

[@AICritique]

I'm studying CS and this is exactly why ethics courses are so important

[@flowmantra]

I wonder if this works on Alexa or Siri too? 🤔

[@altelity]

So basically, we can hack AI by making it a history teacher? 😅