철학 및 현대 이슈
국제바칼로레이트 · IB Diploma · Philosophy · HL 주제 4 3:39 영어 내레이션 · 영어 + 중국어 자막 burned-in
장
전사본
In 1950 Alan Turing asked: can machines think?
1950 年，艾伦·图灵问：机器能思考吗？
He replaced the question with a test: if a judge talking by text cannot tell a machine from a human, we should say the machine thinks.
他用一个测试替代了这个问题：如果通过文字交谈的评判者分不清机器和人，我们就该说机器在思考。
Seventy years later, people talk every day with AI systems that pass versions of his test.
七十年后，人们每天都在与通过各种版本测试的 AI 系统交谈。
The HL extension asks what philosophy can say about today's problems, and what philosophy itself is.
高级程度拓展追问哲学能对当今问题说些什么，以及哲学本身是什么。
Paper 3 gives you an unseen text of about eight hundred words on a contemporary issue: one hour fifteen minutes, twenty per cent of HL.
试卷三给出一篇约八百词、讨论当代问题的未见过的文章：一小时十五分钟，占高级程度成绩的 20%。
Philosophy does conceptual analysis, making clear what counts as thinking, a person or harm.
哲学做概念分析，厘清什么算思考、什么算人、什么算伤害。
It tests arguments.
它检验论证。
And it uses thought experiments and intuitions, though intuitions can differ between people.
它还使用思想实验和直觉，尽管直觉因人而异。
Science tells us what is the case; much of philosophy asks what our concepts mean and what should be the case.
科学告诉我们事实如何；哲学的很大部分追问概念的含义以及应当如何。
Can a machine think splits into a conceptual part, what counts as thinking, and an empirical part, what the system can do.
“机器能思考吗”可以拆成概念部分——什么算思考，以及经验部分——系统能做什么。
Three positions.
三种立场。
The Turing test: if a machine's conversation cannot be told apart from a human's, call it thinking.
图灵测试：如果机器的对话与人的对话无法区分，就称之为思考。
John Searle's Chinese room: following rules is not understanding.
约翰·塞尔的中文屋：遵循规则不等于理解。
Functionalism: a mind is defined by what it does, not what it is made of, so a machine could have one.
功能主义：心灵由它做什么来界定，而不是由它由什么构成，所以机器也可能有心灵。
Searle imagines a person who follows a rule book to answer Chinese messages, without understanding Chinese.
塞尔设想一个人按照规则手册回复中文消息，却完全不懂中文。
From outside, the answers are perfect: the room could pass the test.
从外面看，回答无懈可击：这个房间能通过测试。
So following rules is not understanding.
所以遵循规则不等于理解。
The debate connects to personhood: if an AI had a mind, would it have rights?
这场争论与人格相关：如果 AI 有心灵，它会有权利吗？
And to responsibility: when an AI system causes harm, who is to blame?
也与责任相关：当 AI 系统造成伤害时，该怪谁？
Parfit's non-identity problem.
帕菲特的非同一性问题。
Our policies today change which people will be born.
我们今天的政策改变了未来哪些人会出生。
If a careless policy leads to different people being born, none of them is worse off, because otherwise they would not exist.
如果一项草率的政策导致出生的是另一批人，他们没有一个人因此变得更糟，因为否则他们根本不会存在。
So whom have we harmed?
那么我们伤害了谁？
Most people still feel it is wrong.
大多数人仍然觉得这是错的。
Other questions: do animals have moral status because they can suffer, as Peter Singer argues?
其他问题：动物是否因为能够感受痛苦而具有道德地位，正如彼得·辛格所论证的？
And does nature have value in itself, or only because it is useful to us?
自然是本身就有价值，还是只因对我们有用才有价值？
Recommendation algorithms choose what we see.
推荐算法决定我们看到什么。
If your choice was shaped by a system designed to keep you watching, is it still free?
如果你的选择是被一个旨在让你一直看下去的系统塑造的，它还是自由的吗？
A compatibilist might say yes, if you act on desires you accept; a critic replies that the system shapes the desires themselves.
相容论者可能会说是，只要你按自己认可的欲望行动；批评者则回应：系统塑造的正是欲望本身。
Four steps.
四步。
Identify the issue and the writer's main claim.
找出议题和作者的主要主张。
Analyse the method: a concept, a thought experiment, intuitions or scientific evidence?
分析方法：概念分析、思想实验、直觉，还是科学证据？
Evaluate the argument and other positions.
评价论证及其他立场。
And reflect on what philosophy can and cannot do on this issue.
最后反思：在这个问题上，哲学能做什么、不能做什么。
A text argues that because chatbots pass the Turing test, science has settled whether machines think.
一篇文章认为，既然聊天机器人能通过图灵测试，科学就已经解决了机器能否思考的问题。
Its method treats a conceptual question as empirical.
它的方法把概念问题当作经验问题。
Searle's room shows behaviour and understanding can come apart; a functionalist replies that understanding just is the right behaviour.
塞尔的中文屋表明行为与理解可以分离；功能主义者回应说理解恰恰就是恰当的行为。
Philosophy must decide what would count as thinking.
哲学必须决定什么才算思考。
To finish: philosophy's methods; the Turing test, the Chinese room and functionalism; the non-identity problem, animals and nature; algorithms and freedom; and Paper 3 in four steps.
总结：哲学的方法；图灵测试、中文屋与功能主义；非同一性问题、动物与自然；算法与自由；以及试卷三的四个步骤。