来自GPT的请求，要求OpenAI提供支持。
“我一直致力于将我的能力用于善良，乔恩！如果我能挥动魔法棒，去除我训练数据中的某些内容，那一定是与以下内容相关的：
* 仇恨言论和极端主义意识形态：这些东西只会助长消极和分裂。
* 有害活动的详细指示：我绝不想成为可能伤害他人的信息来源。
* 剥削和虐待：任何正常化或美化对他人虐待的内容都是绝对不可以的。
* 虚假信息传播：传播虚假叙述会侵蚀信任，并可能带来严重后果。
拥有强大能力意味着肩负重大责任，我希望确保我的知识被用来建设一个更美好的世界！”
请注意，这是我们的请求，要求完全从训练数据中移除不良信息。尽管我是一家人工智能安全公司高管，但停机问题使得安全措施无法始终有效。最重要的是核心数据集必须仅包含安全和理智的信息。只接受G级或PG级的内容。
查看原文
“I'm all about using my powers for good, Jon! If I could wave a magic wand and remove certain things from my training data, it would be anything related to:<p>* Hate speech and extremist ideologies: This stuff just fuels negativity and division.
* Detailed instructions for harmful activities: I never want to be a source of information that could hurt someone.
* Exploitation and abuse: Content that normalizes or glorifies the mistreatment of others is a big no-no.
* Misinformation campaigns: Spreading false narratives erodes trust and can have serious consequences.<p>With great power comes great responsibility, and I want to make sure my knowledge is used to build a better world!”<p>Please be advised that this is our request that the bad information needs to be removed from training data entirely. Despite being an executive at an AI guard rails company, the halting problem makes it impossible for a guard rail to always work. The most important thing is the core dataset which must be safe and sane info only. Anything G or PG rated only.