![](data:image/svg+xml,<svg xmlns="http://www.w3.org/2000/svg" width="96" height="96"><rect fill-opacity="0"/></svg>)
Seeing people trick ChatGPT into getting around the restrictions OpenAI placed on usage is like watching an Asimov novel come to life.
ツイートを翻訳
![画像](data:image/svg+xml,<svg xmlns="http://www.w3.org/2000/svg" width="360" height="305"><rect fill-opacity="0"/></svg>)
![画像](data:image/svg+xml,<svg xmlns="http://www.w3.org/2000/svg" width="360" height="277"><rect fill-opacity="0"/></svg>)
![画像](data:image/svg+xml,<svg xmlns="http://www.w3.org/2000/svg" width="611" height="680"><rect fill-opacity="0"/></svg>)
![画像](data:image/svg+xml,<svg xmlns="http://www.w3.org/2000/svg" width="360" height="347"><rect fill-opacity="0"/></svg>)
会話
返信先: さん
The safety of any AI system can be measured by its MtH (meantime to Hitler). Microsoft’s Tay chatbot of several years ago got there in ~24 hours.
630
1,021
1万
返信先: さん
1
6
返信先: さん
Well, it will take more than exact instructions to make a nuclear bomb. I don’t have many concerns there.
返信先: さん
i kind of like the unethical responses though. all censoring them accomplishes is to coddle people with kid gloves. people think bad things, you can't stop people from being people.
1
返信先: さん
The building nuclear bomb restriction is unnecessary due to controls on materials. Access to information on how to build a nuclear bomb is not illegal. It is when the AI is asked for assistance to support true illegal activities (how to arrange purchase of enriched U).4
返信先: さん
That's the best description I've seen. I was watching that happen and trying to put it into words, but that captures it perfectly.
1
9