AP Computer Science Principles · Unit 5 lecture · Part 11 of 11 · 8 min

Unit 5 wrap-upPut it together

Mr. Thompkins
You did it. Unit 5 is a lot of vocabulary, but it all hangs on one question you can ask about any innovation: who does this help, who does it hurt, and how do we keep people safe while we use it?
TopicBig ideaExam move
5.1 EffectsInnovations help and harm, often in ways nobody plannedName who is affected and how
5.2 Digital divideAccess differs by income, place, age and countryName the group that lacks access and why
5.3 BiasPrograms reflect biased data and designWho was in the data vs who it is used on
5.4 CrowdsourcingMany people online contribute input or dataPeople contributing, or computers computing?
5.5 Legal and ethicalCopyright by default; licenses can open work upPermission plus credit
5.6 PrivacyPII, aggregation, data that is hard to deleteSpot the PII and what combining reveals
5.6 SecurityStrong passwords, multifactor, encryption, defensesName the threat, match the defense
Unit 5 on one page, with the move each topic asks for on the exam.

The threads tie together. Data is the center of almost everything here: biased data makes biased programs, aggregated data reveals people, data that's hard to delete outlives its purpose, and data in transit needs encryption. Access is another: the digital divide decides who benefits from an innovation and who shows up in its data. And responsibility runs through it all: programmers should think about unintended uses, reduce bias, respect licenses and protect users' information.

?Wait, which key decrypts a public key message?
The receiver's private key. The sender encrypts with the receiver's public key, and only the receiver's matching private key can decrypt it.
Pause and think
Connect two topics: how could the digital divide make computing bias worse?
I thought about it. Show me
People without internet access don't show up much in the data that apps and models are trained on. A program trained on that data then works less well for them, and decisions built on it can leave them out. Being offline can mean being underrepresented, and that is a source of bias.

Traps that cost points

✕ Calling a harmful effect a bug.
✓ A program can work perfectly and still harm people. Effects are about impact.
✕ Thinking "unintended" means "harmful".
✓ Unintended effects can be good or bad. It means outside the original purpose.
✕ Thinking algorithms are neutral because they are math.
✓ They learn from data and design choices, biases included.
✕ Fixing bias by deleting one sensitive column.
✓ Stand-in clues remain. Test outcomes for each group.
✕ Mixing up crowdsourcing and distributed computing.
✓ People contributing input versus computers running parts of a program.
✕ Thinking anything online is free to use.
✓ Copyright is the default. Look for a license, then give credit.
✕ Counting two passwords as multifactor.
✓ Both are something you know. You need two different categories.
✕ Encrypting with the sender's key, or keeping the public key secret.
✓ Lock with the receiver's public key, unlock with the receiver's private key.
✕ Mixing up phishing and keylogging.
✓ Phishing tricks you into typing somewhere fake. Keylogging records what you type.

Key terms for Unit 5

$\text{Digital divide}$
Differing access to devices and the internetBy income, place, age, country
$\text{Computing bias}$
Programs reflecting human biasThrough data or design
$\text{Citizen science}$
Research with data from the publicVolunteers on their own devices
$\text{Creative Commons}$
Public license for reusing copyrighted workCreator sets the conditions
$\text{PII}$
Personally identifiable informationSSN, age, biometrics, medical, financial
$\text{MFA} = 2+\ \text{categories}$
Multifactor authenticationKnow, have, are
$\text{public locks, private unlocks}$
Public key encryptionUse the receiver's keys. The private key is never shared
$\text{phishing, keylogging, rogue AP, virus}$
The attacks to knowEach has a matching defense

Exam strategy for Unit 5

  1. Unit 5 is 21 to 26 percent of the multiple choice. Five questions sit on one reading passage about a computing innovation; read the questions first, then the passage.
  2. For a passage question, the right answer can be traced to a sentence in the passage. A choice that's true in general but not supported there is a distractor.
  3. Effects: name the group and the specific effect. Unintended means outside the purpose, good or bad.
  4. Data questions: spot the PII, and ask what combining the pieces would reveal.
  5. Bias questions: who was in the data, and who is the program used on?
  6. Security questions: name the threat in one word, then match its defense.
  7. In Section II, Written Response 1 asks about your program's users and purpose, which is impact thinking. Acknowledge any code you didn't write, including AI-generated code.
Mixed practice: a reading passage
A passage describes StudyMatch, an app that pairs students with tutors. Students sign up with their name, school, grade and phone number. The app learns from past ratings which tutors to suggest, and it sends a login code by text message. Most early users came from two large suburban schools. (a) Identify two pieces of PII the app collects. (b) Describe a possible bias and its source. (c) Identify the authentication method and whether it is multifactor. (d) Give one beneficial effect and one effect on the digital divide.
  1. (a) Name and phone number (also school and grade, which help identify a student when combined).
  2. (b) The suggestions are learned from past ratings, mostly from two suburban schools. The app may suggest tutors who suit those students and work less well for students from other schools: bias from unrepresentative training data.
    Why?
    Who was in the data versus who the program is used on: two suburban schools versus everyone.
  3. (c) A code sent by text message proves possession of the phone. If it is the only step, it is single-factor. Paired with a password, it would be multifactor.
    Why?
    Read carefully: the passage mentions only the code. Don't assume a password that isn't there.
  4. (d) Beneficial: students can find help quickly outside school hours. Digital divide: students without a smartphone or reliable service can't sign up or get codes, so they miss the benefit.

Answer: (a) Name and phone number. (b) Suggestions trained mostly on two suburban schools may serve other students worse. (c) A texted code (something you have), single-factor unless paired with something else. (d) Quick help for students; students without phones or service are left out.

Stuck? Get a hint
  1. For (a), which facts point to a specific person?
  2. For (b), compare who the app learned from with who will use it.
  3. For (c), count the categories of evidence the passage actually mentions.
Check yourself
A city launches a free app that reports potholes using photos and GPS from residents' phones. Which TWO ideas from Unit 5 does this most directly involve? Pick the choice that names both.
  1. ACrowdsourcing, and a possible digital divide in which neighborhoods get reported
  2. BPublic key encryption, and Creative Commons licensing
  3. CKeylogging, and a rogue access point
  4. DDistributed computing, and open access research
Show the answer

A. Residents contribute information through the internet, which is crowdsourcing. And neighborhoods with fewer smartphone users would report fewer potholes, so repairs could follow the digital divide. The other pairs don't fit the scenario.

Check yourself
Which action best protects a user if a website they use has its passwords stolen?
  1. AUsing the same strong password on every site
  2. BUsing a unique password for each site and turning on multifactor authentication
  3. CSharing the password only with close friends
  4. DChoosing a password based on a pet's name
Show the answer

B. A unique password keeps one leak from opening other accounts, and multifactor means the stolen password alone isn't enough. Reusing even a strong password, the first choice, turns one leak into many.

Mr. Thompkins
That's Unit 5, and that's all five big ideas. Go take the Unit 5 quiz, and practice reading the scenario before the choices. When a question stings, come back to its section and open the recall cards. You've got this.