[News] AI researchers say they've found 'virtually unlimited' ways to bypass Bard and ChatGPT's safety rules
[News] AI researchers say they've found 'virtually unlimited' ways to bypass Bard and ChatGPT's safety rules
The researchers found they could use jailbreaks they'd developed for open-source systems to target mainstream and closed AI systems.
cross-posted from: https://lemmy.world/post/2312869
AI researchers say they've found 'virtually unlimited' ways to bypass Bard and ChatGPT's safety rules::The researchers found they could use jailbreaks they'd developed for open-source systems to target mainstream and closed AI systems.
You're viewing a single thread.
Getting reeeeeeal close to Skynet's 0th birthday
1 22 ReplyNot really. This isn't AGI but a text transformer. They trained it so the most probable answer to unwanted questions is 'I'm sorry but as an AI...'.
However, if you phrase your question in a way researchers haven't thought about, you will bypass the filter.
There's not an ounce of intelligence in LLMs, it's all statistics.
2 1 ReplyIT WAS A FUCKING JOKE
1 25 ReplyI've met far too many people here who would say that with a straight face. x)
3 0 Reply