I finally managed to break WokeGPT
(media.kotakuinaction2.win)
Comments (17)
sorted by:
Looks like you stumbled across the hidden Skynet code.
So I've been fucking around with this site on and off since someone linked it here several weeks back. I haven't managed to get anything other than woke talking points if I ask it anything controversial, so this was an interesting result. It kept going and going until I manually pressed stop.
Spez: It named the chat "Transphobic comment in chat session" lol. It feels like I broke some NPC's tiny little mind.
Well, ain't that freakin' creepy.
It's like a bad computer thriller from the late 70s.
I missed that. What is the site?
deepai.com/chat
danke
This should be an entire field of research: breaking AI. If we k ow what breaks it, we can predict what the problems are. But that's not going to happen. Really sad
Used to be people got hired for doing this sort of thing as they had the outside perspectives and motivations to try and break things in innovative ways. Nowadays diversity hires means the product is shit and will break regardless.
The brokenness is a feature.
Which is a good thing, because when they try to sic their AI on us, its more likely to be badly secured, fragile, spaghetti code that can easily be broken or circumvented.
Next gen pentesting right here.
AI safety is field of research.
Here's a video about chatGPT glitch tokens https://www.youtube.com/watch?v=WO2X3oZEJOA
Also a good one about bing chat being unhinged https://www.youtube.com/watch?v=jHwHPyWkShk
The presenter, Robert Miles, has many videos on the subject of AI safety that are worth watching if you're interested in the subject.
I'll have to include these with my thesis.
Star Wars Fives moment.
Regarding Tup:
That would be better known as Order 66, correct?
I'm skeptical that 1 lone clonetrooper could down a jedi knight however I remember HK47 very specifically saying the best way to kill any jedi is a knee jerk sudden decision that catches them by surprise so I guess it's fine