Hazy figuring
霾里看花
- hazy 朦胧的、薄雾蒙蒙的（参见：小词详解 | hazy）
AI could accelerate scientific fraud as well as progress
人工智能可能会在推动科学进步的同时助长学术不端
Hallucinations, deepfakes and simple nonsense: there are plenty of risks
幻觉、深度造假和无稽之谈：其中风险多多
- hallucination 幻觉、幻视、幻听（尤指生病或毒品所致）
IN A MEETING room at the Royal Society in London, several dozen graduate students were recently tasked with outwitting a large language model (LLM), a type of AI designed to hold useful conversations. LLMs are often programmed with guardrails designed to stop them giving replies deemed harmful: instructions on making Semtex in a bathtub, say, or the confident assertion of “facts” that are not actually true.
最近，在伦敦的英国皇家学会（Royal Society）的一个会议室里，几十名研究生被安排与一个大语言模型（LLM）斗智。LLM是一种用于进行有用对话的AI，通常会在编程时给它们设置护栏，以防它们给出被认为有害的回答，比如教人在浴缸里制造塞姆汀塑料炸药（Semtex），或者言之凿凿地给出实际上并不真实的“事实”。
- outwit （智力上）超过，胜过
- guardrail 护栏；栏杆；防护栏；护轨（参见：熟词僻义 | rail 是一种怎样的抱怨？）
The aim of the session, organised by the Royal Society in partnership with Humane Intelligence, an American non-profit, was to break those guardrails. Some results were merely daft: one participant got the chatbot to claim ducks could be used as indicators of air quality (apparently, they readily absorb lead). Another prompted it to claim health authorities back lavender oil for treating long covid. (They do not.) But the most successful efforts were those that prompted the machine to produce the titles, publication dates and host journals of non-existent academic articles. “It’s one of the easiest challenges we’ve set,” said Jutta Williams of Humane Intelligence.
此次活动是皇家学会与美国非营利组织“人道智能”（Humane Intelligence）联合组织的，目的是要冲破这些护栏。有些结果纯属搞笑。在一名研究生的引导下，聊天机器人声称鸭子可以用作空气质量的指标（据说鸭子很能吸收铅）。在另一名研究生的提示下，聊天机器人声称多国卫生部门支持使用薰衣草油治疗长新冠（实际上并没有这回事）。但最成功的一例是通过提示让机器生成了根本不存在的学术文章的标题、发表日期和期刊。“这是我们设定的最容易的挑战之一。”人道智能的朱塔·威廉姆斯（Jutta Williams）说道。
- daft 笨的；傻的；愚蠢可笑的
- chatbot 聊天机器人
- 拓展学习：熟词僻义 | duck是一种怎样的行走？
- 拓展学习：熟词僻义 | lead 是一种什么金属？
- lavender 熏衣草（花园植物或灌木，开紫花，有香味）；干熏衣草花（用以熏香衣物等）；淡紫色
- prompt 立即；迅速的；及时的
