当前关于AI安全的讨论不成比例地强调假设的未来风险，而非紧迫的、切实的现有问题。迫切需要更多工程师专注于解决现有问题，例如有偏见的输出、不可靠的应用程序编程接口和糟糕的文档。 AI
影响 强调了AI发展中的一个关键差距，敦促将重点从推测的未来风险转向实际问题的解决。
排序理由 一位知名人士发表的关于AI安全优先事项的观点文章。
在 Mastodon — mastodon.social 阅读 →
AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →
AI安全讨论忽略当前问题，需要更多实践工程
报道来源 [1]
-
Been thinking about how much of "AI safety" discourse focuses on hypothetical future risks while ignoring real present-day problems — biased outputs, API reliab
Been thinking about how much of "AI safety" discourse focuses on hypothetical future risks while ignoring real present-day problems — biased outputs, API reliability, bad documentation. We need more engineers fixing what is actually broken today. # AI # engineering # practicality