Chapter 14 — Privacy foundations | Bachelor's Quantum and post-quantum security

Open, free learning for anyone who wants to protect people and technology. No account is required to read this text or its worked answers.

Chapter 14 — Privacy foundations

Lesson map

Lesson 14 published. Map, example, EX027.

flowchart TD
  N0["Learning outcomes"]
  N1["14.1 Principles"]
  N2["14.2 Lifetime of harm"]
  N3["14.3 Minimization as a quantum control"]
  N4["14.5 Practical activity"]
  N5["Synthesis"]
  N0 --> N1 --> N2 --> N3 --> N4 --> N5
  N5 --> EX["Worked example"]
  EX --> Q["Exercise EX027"]

Logical drawing

flowchart LR
  N0["Learning outcomes"]
  N1["14.1 Principles"]
  N2["14.2 Lifetime of harm"]
  N3["14.3 Minimization as a quantum control"]
  N4["14.5 Practical activity"]
  N5["Synthesis"]
  N0 --> N1 --> N2 --> N3 --> N4 --> N5 --> Q["EX027"]

[Learning outcomes]
    |
    v
[14.1 Principles]
    |
    v
[14.2 Lifetime of harm]
    |
    v
[14.3 Minimization as a quantum control]
    |
    v
[14.5 Practical activity]
    |
    v
[Synthesis]
    |
    v
[Worked example] --> [Exercise EX027]

Learning outcomes

You will (1) state core privacy principles such as purpose limitation and minimization; (2) relate privacy harm to confidentiality lifetime; (3) explain why harvest-now-decrypt-later is a privacy problem; and (4) design data-minimizing measurement for migration work.

14.1 Principles

Privacy law and practice share common principles: collect personal data for a stated purpose, collect no more than needed (minimization), keep it no longer than needed (storage limitation), protect it appropriately, and be transparent. These principles are about people, not systems; a privacy harm is a harm to a member whose data is exposed, even if the organization suffers no loss.

For a quantum migration these principles shape two things: which data most needs protection, and how the migration work itself handles personal information found along the way.

14.2 Lifetime of harm

Some personal data stops mattering quickly; other data stays sensitive for a lifetime — health conditions, financial distress, identity documents. Harvest-now-decrypt-later turns long-lived sensitivity into present risk: traffic recorded today can reveal tomorrow what a member disclosed in confidence decades earlier. For privacy, the relevant x in Mosca's inequality is the lifetime of the harm to the person, which can exceed the organization's retention period. Deleting a record from HCU's database does not delete a copy an adversary recorded in transit.

14.3 Minimization as a quantum control

The cheapest protection for data is not to send it. Reducing what crosses networks — tokenising account numbers, keeping sensitive fields at the data centre and sending only derived results — lowers HNDL exposure regardless of algorithm. Minimization also applies to migration measurement itself: an inventory needs algorithm names and host identifiers, not message contents.

HCU's first review of branch traffic found member names travelling with transactions that needed only an account token. Replacing the names with tokens reduced the exposure at once, before any algorithm changed.

14.4 Worked example

HCU's loan application flow sends full applicant profiles from branches to head office, including medical notes for hardship cases. A privacy review finds that underwriters at head office need only a hardship flag and an amount. Changing the form so that medical notes stay in the branch case file reduces both the privacy footprint and the data exposed to future decryption, before any post-quantum algorithm is deployed. Hybrid key exchange on the branch VPN is then scheduled for the remaining flows.

14.5 Practical activity

For three synthetic HCU data flows, list the personal data fields, the purpose, and which fields could be removed, tokenised or aggregated before transmission.

Typical candidates are full names sent where an identifier would do, dates of birth sent for checks that only need an age band, and account numbers sent where a token works. For each field you remove, say who would notice and whether any process breaks. Then estimate how much the change reduces the long-lived data crossing the network, which is the exposure harvest now, decrypt later depends on.

14.6 Common errors and limits

Equating retention period with lifetime of harm.
Collecting full packet captures for an inventory.
Treating PQC as a substitute for minimization rather than a complement.

The harm from disclosure can last far longer than the record is kept: a deleted mortgage file that was recorded in transit can still be decrypted later. Full captures collect personal data the inventory never needed and create a new store to protect. Post-quantum algorithms protect what is sent; minimization reduces what is sent at all. The two work together, and neither replaces the other.

Exercises

EX027 — Lifetime of harm. Give two data types whose privacy harm outlasts HCU's retention period and explain the consequence for HNDL. Rubric: 3 = two examples, reasoning linked to recorded traffic.
EX028 — Minimize the inventory. Specify the minimum fields a passive TLS inventory needs and two fields it should not collect. Rubric: 3 = justified lists.

Synthesis

Privacy harm can last longer than any retention schedule, which makes harvest-now-decrypt-later a privacy issue. Minimization reduces exposure under every algorithm and should precede and accompany PQC.

Continue studying — chapter 14 of 40

← Chapter 13 (draft) · Course chapter index · Chapter 15 (draft) →