AP CSP Privacy PII
AP CSP Privacy & PII: Complete Guide (2025‑2026)
Personally Identifiable Information (PII) is any data that can identify an individual — alone or in combination. The subtle trap AP CSP tests: data that looks anonymous often is not. Zip code + birthdate + gender alone re-identifies most people. Aggregation turns harmless fields into identifying combinations. Metadata reveals as much as content. The exam tests whether you can spot privacy risks that are not obvious on the surface.
Contents
How Aggregation Creates Identity Risk
Each field below seems harmless in isolation. Together, they uniquely identify a person even without a name.
Removing a name is not anonymization. Aggregation of remaining fields creates a quasi-identifier that re-identifies most individuals.
A hospital publishes a “de-identified” patient dataset. They removed names and Social Security numbers. The remaining fields include age, ZIP code, diagnosis code, admission date, and insurance provider. A researcher claims the data is now fully anonymous and safe to publish publicly.
What is wrong with the researcher’s claim? What specific risk remains?
The data is pseudonymized, not anonymized. The remaining fields — especially age + ZIP + diagnosis + admission date — form a quasi-identifier cluster that can re-identify specific patients, particularly those with rare diagnoses or in small geographic areas. Removing names and SSNs is necessary but not sufficient for true anonymization.
Metadata: More Revealing Than Content
Metadata is “data about data.” When you send a message, metadata records who you contacted, when, for how long, and from where — all without reading a single word of content.
- “[encrypted — unreadable]”
- “[encrypted — unreadable]”
- “[encrypted — unreadable]”
- HTTPS protects this layer
- Sender: patient_A → oncology clinic
- Duration: 47 minutes
- Time: Tuesday 9:03 AM
- Location: home address
A messaging app encrypts all message content end-to-end. The government subpoenas the company’s metadata logs: sender, recipient, timestamps, and call durations. A user argues: “My messages are encrypted, so my privacy is protected.”
Is the user’s argument correct? What does the metadata reveal?
Incorrect. End-to-end encryption protects message content, not metadata. The metadata logs reveal communication patterns: who the user contacts, how often, for how long, and at what times. In the example above, a patient contacting an oncology clinic for 47 minutes reveals likely diagnosis information without decrypting any message.
Privacy vs. Security Trade-offs
- Users control their own information
- Data not retained after purpose served
- Opt-in consent before collection
- Anonymization before analysis
- Behavior logs identify attackers
- Long retention enables forensics
- Broad collection catches anomalies
- Identifiable records enable accountability
A city deploys smart streetlights that record pedestrian movement 24/7 to optimize traffic signals and detect accidents. The data is stored for 90 days. Officials argue this improves public safety. Privacy advocates argue it creates a surveillance infrastructure.
What is the core privacy trade-off here? Is there a way to reduce privacy cost while keeping the safety benefit?
The trade-off: public safety benefit (accident detection, traffic optimization) vs. creation of granular movement records for every citizen. Privacy-preserving alternatives include: aggregating pedestrian counts without storing paths, using edge processing that never transmits raw video to central servers, or reducing retention to 24-48 hours. The AP exam tests your ability to identify trade-offs and evaluate whether the data collected is proportional to the stated purpose.
Common Exam Pitfalls
Removing names is necessary but not sufficient. Aggregation of remaining fields frequently re-identifies individuals. The AP exam tests aggregation risk specifically.
Encryption protects data in transit from eavesdroppers. It does not prevent the collecting party from using, analyzing, or selling your data. HTTPS means your ISP can’t read it — not that the website can’t.
Metadata (who, when, where, how long) frequently reveals more than content. Courts and intelligence agencies prioritize metadata access precisely because of its revealing nature.
On most platforms, opt-out of targeted advertising means data is collected but not used for that specific purpose. Collection and use are separate. Read privacy policies carefully — opt-out rarely means no collection.
Check for Understanding
1. A social network stores: username, city, employer, university attended, and graduation year. A user requests their username be removed to protect privacy. Which statement is most accurate?
- The user is now fully de-identified because the unique identifier (username) was removed.
- The combination of city, employer, university, and graduation year may still uniquely identify the user.
- Removing the username satisfies all legal definitions of anonymization in all jurisdictions.
- The remaining fields contain no PII because none is a government-issued identifier.
2. A messaging app encrypts all message content end-to-end. Which of the following is still vulnerable despite this encryption?
- The content of messages between users
- The encryption key used to secure messages
- Metadata logs recording who communicates with whom, when, and for how long
- The app’s source code on the user’s device
3. Consider these statements about PII:
I. Direct PII includes names, SSNs, and email addresses.
II. Indirect PII can only become identifying if combined with direct PII.
III. An IP address can qualify as PII in some privacy frameworks.
Which statements are correct?
- I only
- I and II only
- I and III only
- I, II, and III
4. A fitness app collects step counts and sleep patterns. It shares this data with “research partners” after removing usernames. This practice is best described as:
- Fully anonymized data sharing because usernames are absent.
- Pseudonymization that retains re-identification risk through behavioral fingerprinting.
- A secure and privacy-compliant practice because biometric data is inherently imprecise.
- Illegal under all modern privacy regulations.
5. Which scenario best illustrates the secondary use problem in data privacy?
- A website uses HTTPS to encrypt login credentials during transmission.
- A city installs traffic sensors for congestion pricing; law enforcement later uses the location data to track individuals.
- A company stores passwords using bcrypt hashing.
- A social network allows users to download a copy of their data.
6. A student argues: ‘If I use a VPN, websites cannot collect any personal data about me.’ This statement is:
- Correct — VPNs encrypt all traffic, preventing all data collection.
- Correct — VPNs mask IP addresses, which are the only form of PII websites collect.
- Incorrect — websites collect data through cookies, login sessions, and fingerprinting that persist through a VPN.
- Incorrect only if the VPN provider is located in the same country as the user.
Frequently Asked Questions
How the AP Exam Tests This
- Identify whether specific data (alone or combined) constitutes PII
- Explain the concept of data aggregation and why combining non-PII can create PII
- Describe a specific privacy risk created by a computing technology
- I/II/III: which statements about privacy and PII are correct
- Identify which data collection practice violates a reasonable expectation of privacy
7. A website collects: browsing history, ZIP code, age range, and device type — none of which are PII individually. Why might the combination be a privacy concern?
- None of these data points are ever sensitive.
- Combining multiple data points can uniquely identify an individual even when no single point is PII.
- The combination is only a concern if names are included.
- Data aggregation never creates privacy risks.
8. A person posts photos on social media with location tagging enabled. Which privacy risk does this most directly create?
- Their password can be extracted from the photos.
- Their physical location and daily patterns can be inferred from the metadata.
- Their social media account becomes vulnerable to hacking.
- Their photos can be used for facial recognition only.
9. Consider: I. Data that is not PII alone can become PII when combined with other data. II. Encryption guarantees complete privacy of personal data. III. Users giving consent to data collection are always fully informed of how data will be used.
- I only
- I and II only
- I, II, and III
- II and III only
10. A company sells its user database to a third party after users agreed to a terms-of-service that included sharing data with ‘trusted partners.’ Users did not read this clause. This situation illustrates:
- No privacy issue — users gave legal consent.
- A tension between legal consent and meaningful informed consent, raising privacy and ethical concerns.
- Illegal data sale regardless of terms of service.
- A technical privacy breach.
Get in Touch
Whether you're a student, parent, or teacher — I'd love to hear from you.
Just want free AP CS resources?
Enter your email below and check the subscribe box — no message needed. Students get daily practice questions and study tips. Teachers get curriculum resources and teaching strategies.
Message Sent!
Thanks for reaching out. I'll get back to you within 24 hours.
Prefer email? Reach me directly at [email protected]