Impact of Computing
AP Computer Science Principles Topic 5 8:12 English narration · English + 中文 subtitles burned in
Chapters
Transcript
In nineteen eighty-nine, one scientist built a small tool so physicists could swap papers faster.
1989年，一位科学家做了一个小工具，只是想让物理学家更快地交换论文。
He was not planning any of this.
他并没有计划过后来发生的一切。
A car that drives itself.
一辆自己驾驶的汽车。
A wall of cameras watching a street.
一整面监视街道的摄像头墙。
A mountain of dead machines.
一座报废机器堆成的小山。
Each one helps somebody and hurts somebody else — and almost none of it was the plan.
它们每一个都在帮助某些人，也在伤害另一些人—— 而这几乎都不在计划之内。
This unit is the impact of computing, and it comes in six parts. Beneficial and harmful effects. The digital divide. Computing bias. Crowdsourcing. Legal and ethical concerns. And safe computing — how you protect your own data. All six are examined. Let's begin.
本单元讲计算的影响，一共分六个部分：有益与有害的影响；数字鸿沟；算法偏见； 众包；法律与伦理问题；以及安全计算，也就是如何保护你自己的数据。 这六部分都会考。 让我们开始吧。
Here is the first big idea.
第一个核心观点。
One effect can be beneficial and harmful at the same time, and often to different people at once: a ride-share app helps riders and hurts taxi drivers.
同一个影响可以同时既有益又有害，而且常常是对不同的人同时发生： 一个网约车应用帮助了乘客，却伤害了出租车司机。
So weigh both sides, never one.
所以两面都要权衡，不能只看一面。
An innovation can also bring unintended consequences beyond its purpose — a map app made to guide you also records everywhere you go.
一项创新还可能带来意外后果，超出它原本的目的—— 一个用来指路的地图应用，也在记录你去过的每一个地方。
So how do you judge an innovation?
那么，怎样评价一项创新？
Point it at the public.
把它对准公众。
Ask about health and safety: does it keep people safer, or put them at risk?
先问健康与安全：它让人更安全， 还是让人更危险？
Ask about the public interest: who gains, and who loses?
再问公共利益：谁得益，谁受损？
Then name one clear benefit to the public, and one real concern of the public.
然后说出一个对公众明确的好处， 和一个公众真实的担忧。
Four questions, one balanced answer.
四个问题，一个两面兼顾的答案。
The digital divide means unequal access to computers and the internet.
数字鸿沟指的是使用电脑和互联网的机会并不平等。
Three things drive it.
造成它的原因有三个。
Money: a device and a data plan both cost money.
钱：一台设备和一份流量套餐都要花钱。
Where you live: a city gets fast internet years before a village.
住在哪里：城市比乡村早很多年用上高速网络。
Age, education and language decide who even tries.
年龄、教育和语言决定了谁会去尝试。
With access come school, jobs and health care. Without it, a region falls further behind.
有了网络，就有学习、工作和医疗； 没有网络，一个地区就会越落越远。
People, companies and governments widen that gap, or close it.
个人、企业和政府既可能扩大这道鸿沟，也可能把它填平。
This is what the divide looks like.
数字鸿沟看起来就是这样。
A public library, where one shared row of computers is the only way online for a whole district.
一间公共图书馆，那一排共用的电脑是整个地区唯一的上网方式。
A satellite dish on a cabin, because the cable never reached that valley.
一口装在小屋上的卫星天线，因为电缆从未通到那个山谷。
And hold on to one thing: no way online at all is the divide, and the divide is about access.
记住一点：完全没有办法上网，这叫鸿沟，鸿沟讲的是接入。
Fairness once you are online is a different word — it is next.
而上网之后是否被公平对待，那是另一个词——下面就讲它。
Computing bias — also called algorithmic bias.
算法偏见——也叫 algorithmic bias。
A system is biased when it treats one group unfairly.
当一个系统对某一群体不公平时，我们说它有偏见。
Two sources.
它有两个来源。
The data: a model learned from the past repeats it, unfair decisions included.
数据：从过去学来的模型会照搬过去，连不公平的决定也一起搬过来。
The design: the team who built it chose who the typical user is, so everyone else does worse.
设计：开发团队选定了谁是典型用户，其他人得到的结果就更差。
Bias enters at every stage — collecting, coding, testing, selling.
偏见会出现在每一个环节——收集、编码、测试、销售。
It is almost always an accident, and that does not make it fair.
它几乎总是无心之失，但这并不能让它变得公平。
So test the results for every group, then fix the data.
所以要对每个群体分别检验结果，再修正数据。
A hiring tool is trained on ten years of a company's past hires, mostly one group, and it now rejects strong applicants from other groups.
一个招聘工具用公司十年来的录用记录训练，而这些人大多来自同一个群体。 现在，它拒绝了其他群体中很优秀的申请者。
Name the problem, its cause, and a fix. Pause here and try.
请说出问题、原因，以及一个改进办法。
The problem is computing bias: good people are rejected unfairly.
先暂停，自己试一试。 问题是算法偏见：合格的人被不公平地拒绝了。
The cause is the training data, not the code — it copied a historical pattern.
原因在训练数据，而不在代码——这个工具照抄了历史上的模式。
One fix is to test the results group by group, then retrain on fairer data.
一个改进办法是分组检验结果，然后用更公平的数据重新训练。
A full answer states the harm and the cause.
满分答案要把危害和原因两样都写出来。
Crowdsourcing means getting input, work or money from a large group of people over the internet.
众包指的是通过互联网，从一大群人那里获得点子、劳动或资金。
These volunteers are writing one shared encyclopedia.
画面里的志愿者正在共同编写同一部百科全书。
Citizen science is the same idea in research: ordinary people, not scientists, collect real data on their own devices.
公民科学是研究领域里的同一种做法：普通人——不是科学家——用自己的设备收集真实数据。
Crowdfunding is the same idea with money — many small payments start a project no bank would fund.
众筹是资金上的同一种做法——很多人的小额付款，能启动一个银行不愿投资的项目。
Law and ethics.
法律与伦理。
What you make on a computer is your intellectual property, and copyright protects creators' work — copying it needs permission or a license.
你在电脑上创作的东西就是你的知识产权，版权保护创作者的作品—— 别人要复制需要许可或授权。
Plagiarism is using someone else's work as your own — unethical, often illegal, so cite every source.
抄袭就是把别人的作品当作自己的——既不道德， 也常常违法，所以每次都要注明来源。
Legal ways to share exist: Creative Commons, open source and open access.
合法的共享方式是有的：知识共享、开源和开放获取。
Collecting personal data raises privacy questions.
收集个人数据会带来隐私问题。
And just because something is possible, that does not make it legal.
而技术上做得到，并不等于合法。
Personally identifiable information — PII — is anything that identifies you: your name, age, phone number, address, even a fingerprint.
个人身份信息——也就是 PII——是任何能识别出你的信息： 姓名、年龄、电话号码、住址，甚至指纹。
Share it carefully, because it can be misused for identity theft.
要谨慎分享，因为它可能被滥用于身份盗窃。
Most of it you never type in.
其中大部分你从来没有主动填写过。
A search engine keeps what you searched for. A website records which pages you opened. Your phone reports where it has been.
搜索引擎会保存你搜索过的内容； 网站会记录你打开过哪些页面；手机会上报它去过的地方。
Put those pieces together and they reveal things you never posted.
把这些碎片拼起来，就能看出你从未发布过的事情。
Once online, it is very hard to delete.
信息一旦上网，就很难删除。
The most common attack does not break the computer at all — it tricks the person.
最常见的攻击根本不是攻破电脑，而是骗过人。
This is phishing.
这就是网络钓鱼。
A message pretends to be your bank and says something urgent: your account will close in one hour.
一条消息假装来自你的银行，说了一件很急的事：你的账户一小时后就会关闭。
The link opens a page that looks exactly right, and the password you type goes straight to the attacker.
链接打开的页面看起来完全正常，而你输入的密码，直接送到了攻击者手里。
Phishing is one of four ways in, and the exam names all four.
入侵途径共有四种，考试都会点名，网络钓鱼只是其中之一。
Here is another: a rogue access point.
这是另一种：恶意接入点。
In a café, two wireless networks carry the same name.
在咖啡馆里，有两个同名的无线网络。
One is real; the fake one belongs to an attacker, and your laptop cannot tell them apart.
一个是真的；那个假的属于攻击者， 而你的笔记本电脑分辨不出来。
Connect to the fake, and anything you send without encryption can be read.
一旦连上假的，你发出的一切未加密内容都能被读取。
The last two are keylogging, which records every key you press, and an untrusted download, which can carry malware.
剩下两种是键盘记录——它会记录你按下的每一个键—— 以及来路不明的免费下载，它可能带着恶意软件。
Five defences, in the exam's words.
五道防线，都用考试要求的说法。
A strong password: hard to guess, different on every site.
强密码：别人难猜，而且每个网站都不一样。
Multi-factor authentication: two steps of different kinds — something you know, have, or are.
多因素认证：两个步骤要来自不同类别——你知道的、你拥有的、你本身的。
Encryption, which scrambles data so only the key reads it.
加密，它把数据打乱，只有拿到密钥才能读懂。
Software updates, which close the flaws attackers use.
软件更新，用来堵住攻击者利用的漏洞。
And check what each app is allowed to collect.
还有，检查每个应用被允许收集什么。
No single defence is enough, so use layers.
没有哪一道防线是足够的，所以要多层设防。
Encryption deserves a closer look, because it guards data while it travels.
加密值得再看仔细一点，因为它正是在数据传输途中保护数据的工具。
Your readable message is called plaintext.
你能读懂的原始消息叫做明文。
An encryption algorithm scrambles it using a key, and what comes out is cipher text, which means nothing to anyone who intercepts it.
加密算法用一把密钥把它打乱，输出的东西叫做密文， 对半路截获它的人来说毫无意义。
At the far end a decryption algorithm, with the matching key, turns it back into the original plaintext.
到了另一端，解密算法配上相应的密钥， 再把它还原成最初的明文。
There are two families.
加密有两大类。
In symmetric key encryption, one shared secret key both locks and unlocks the message, so both people must already have it.
在对称密钥加密中，一把共享的密钥既能上锁也能解锁， 所以双方必须事先都拿到它。
In public key encryption there are two keys: a public key anyone may use to lock a message, and a private key, kept secret, that alone can open it.
在公钥加密中有两把钥匙：一把公钥， 任何人都可以用它把消息锁上；一把私钥，必须保密，只有它才能打开。
Now a full written response.
现在来写一份完整的文字作答。
A fitness band records your heart rate and location, and posts a weekly summary to friends.
一个运动手环记录你的心率和位置， 并每周把摘要发给朋友。
Give one benefit, one harm, one privacy risk, and one practice that reduces it.
请给出一个好处、一个坏处、一个隐私风险， 以及一个能降低该风险的做法。
Pause and plan four sentences.
先暂停，把四句话想好。
The benefit: an unsafe heart rate is spotted at once.
好处：心率不正常会被立刻发现。
The harm: friends learn your routine.
坏处：朋友能看出你的作息。
The privacy risk: location is personally identifiable information.
隐私风险：位置属于个人身份信息。
The practice: turn location sharing off.
做法：关闭位置共享。
Tie every point to this band, not to technology in general.
每一点都要扣住这个手环来写，而不是泛泛地谈技术。
Three marks students throw away.
三个学生常丢的分。
First, argue both sides: if a question asks about the effects of an innovation, give a benefit and a harm.
第一，两面都要写：如果题目问一项创新的影响， 就要给出一个好处和一个坏处。
Second, do not swap the two big words: the divide is about access, bias is about fairness.
第二，不要把两个关键词弄混：鸿沟讲接入，偏见讲公平。
Use correct terms for data concerns: PII, privacy, security, and algorithmic bias.
数据相关用语要说对：PII、隐私、安全和算法偏见。
Third, name the innovation and name the effect; technology in general earns nothing.
第三，写出具体的创新和具体的影响； 泛泛地谈技术是拿不到分的。
Get those right, and this unit is yours.
做到这三点，这个单元就是你的了。