Cafe Conclave III

How to Have a Healthy Relationship with AI

How much do you all trust AI?

AI dependency has become a concern recently, but I think there's no real problem with using AI as long as you don't stop using it entirely and instead make sure you're using it properly.

On a slightly different note, I've also been hearing a lot about AI being involved in recent cybercrime. Regulations are supposed to be getting stricter, but can DAN still be used nowadays?

↑ 2

讨论

评论

1

每一层先显示最新的评论;回复仍在对应评论下方。

写信 加入后即可参与讨论。

The trust question is really about verification, not the tool. If you can't check the output yourself, you're not using it properly regardless of how you prompt it. On crime use: models help with the boring parts like writing and sorting, the actual capability was always in the person holding it. Jailbreaks come and go with every model update, so treating DAN as a stable workaround is optimistic.
↑ 0 回复