AI Basics 人工智能基础
What does AI stand for? - AI 代表什么？
Artificial Intelligence - 人工智能
What is a prompt? - 什么是提示词（Prompt）？
An instruction or request given to AI - 给 AI 的指令或请求
What should you do before immediately accepting an AI answer? - 在立即接受 AI 的答案之前，你应该做什么？
Check it - 检查它
What does risk mean when using AI? - 使用 AI 时，“风险”是什么意思？
The possibility that something harmful or unwanted could happen - 可能发生有害或不希望发生事情的可能性
What does bias mean? - 什么是偏见（Bias）？
Information or decisions being unfairly influenced in a particular direction - 信息或决策受到某一方向的不公平影响
What does AI use to find patterns and make predictions or responses? - AI 使用什么来寻找模式并作出预测或回应？
Data - 数据
What is the process of writing clear and effective prompts called? - 编写清晰且有效的提示词的过程叫什么？
Prompt engineering - 提示工程
What is an AI answer that sounds convincing but is incorrect called? - 一个听起来很有说服力但实际上错误的 AI 答案叫什么？
Hallucination / confabulation - 幻觉 / 虚构
Who should remain responsible for the final decision when using AI? - 使用 AI 时，谁应该对最终决定负责？
The human - 人类
What type of bias happens when data is incomplete or unbalanced? - 当数据不完整或不平衡时，会产生什么类型的偏见？
Data / statistical bias - 数据偏见 / 统计偏见
What is AI designed to perform that normally requires human intelligence? - AI 被设计用来执行通常需要人类智能才能完成的什么？
Tasks - 任务
What does the G in GCCF stand for? - GCCF 中的 G 代表什么？
Goal - 目标
What process involves checking whether information from AI is correct? - 检查 AI 提供的信息是否正确的过程叫什么？
Verification - 验证
What should increase when the consequences of an AI decision become greater? - 当 AI 决策的后果变得更加严重时，什么应该增加？
Human judgement and verification - 人类判断和验证
What type of bias can occur when data reflects human attitudes, stereotypes or inequalities? - 当数据反映人类态度、刻板印象或不平等时，可能产生什么类型的偏见？
Societal / human bias - 社会偏见 / 人类偏见
What process allows AI to identify patterns in information? - 哪个过程让 AI 能够识别信息中的模式？
Pattern recognition - 模式识别
What does the C in GCCF stand for? - GCCF 中的 C 代表什么？
Context - 背景
What should you compare an important AI claim with to check whether it is correct? - 为了检查重要的 AI 说法是否正确，你应该把它与什么进行比较？
Reliable evidence / sources - 可靠的证据 / 来源
Why should humans check important AI decisions instead of simply accepting them? - 为什么人类应该检查重要的 AI 决策，而不是直接接受它们？
Because AI can be incorrect and the consequences may be serious - 因为 AI 可能出错，而错误的后果可能很严重
Why can biased data lead to biased AI results? - 为什么有偏见的数据可能导致有偏见的 AI 结果？
Because AI learns patterns from data - 因为 AI 会从数据中学习模式
What can AI produce that may look intelligent even though AI does not think like a human? - AI 可以产生什么，看起来很聪明，但 AI 实际上并不像人类一样思考？
Intelligent-looking results - 看起来很智能的结果
What are the four parts of the GCCF prompt structure? - GCCF 提示词结构的四个部分是什么？
Goal, Context, Constraints, Format - 目标、背景、限制条件、格式
What three-step process does the course recommend before accepting an AI response? - 课程建议在接受 AI 回应之前采用哪三个步骤？
STOP → CHECK → THINK - 停止 → 检查 → 思考
What principle does the course give about consequences, human judgement and verification? - 课程对于后果、人类判断和验证提出了什么原则？
The greater the consequences, the greater the need for human judgement and verification - 后果越严重，就越需要人类进行判断和验证
What are the two main types of bias identified in the course? - 课程中确定的两种主要偏见是什么？
Data/statistical bias and societal/human bias - 数据/统计偏见和社会/人类偏见