CRYPTOGRAPHY EDUCATION · PART 2
Cryptography education for children is built almost entirely on confidentiality. Integrity, authentication and credentials are arriving in children's lives anyway, through policy rather than pedagogy.
Cryptography education, part 2. The first article argued that children can learn the questions before the mathematics. This one follows integrity, authentication and credentials into the classroom and into age-assurance systems.
The first article in this series ended on a table. Four questions a child can ask, four cryptographic properties:
| Question | Property |
|---|---|
| Can somebody else read this? | Confidentiality |
| Has this been changed since it was sent? | Integrity |
| Who produced or authorised it? | Authentication |
| Can somebody later show what was approved? | Evidence and accountability |
Most of the programmes surveyed in the first article enter through the first row. This article is about the other three, and about a timing problem: systems that depend on those properties are beginning to shape children's digital lives faster than educational material is explaining what those systems do.
The card trick has the wrong adversary
Integrity does have a canonical classroom activity, and it is a good one. CS Unplugged's Error Detection activity is a magic trick: a volunteer lays out a grid of two-sided cards, the demonstrator adds a row and a column of extra cards, somebody flips one card while the demonstrator looks away, and the demonstrator finds it. The extra cards are parity bits. The trick works because flipping any single card breaks the parity of exactly one row and exactly one column, and the intersection names the culprit. The activity generalises naturally to ISBN and barcode check digits, which is where children meet the same idea in the physical world.
It is one of the best-designed activities in early computing education. It is also, read as a security lesson, quietly teaching the wrong thing.
The adversary in the card trick is noise. A cosmic ray, a scratched disc, a dropped bit on a wire. Noise flips one card at random, and parity is exactly strong enough to catch that. But the trick's own teaching notes acknowledge the limit: flip four cards arranged at the corners of a rectangle and every row and column still has even parity. The error is invisible.
Random noise almost never produces that pattern. An opponent produces it on the first try.
And there is a deeper problem than the rectangle. Suppose the child doing the flipping knows the rule. Then she does not need a clever pattern at all. She changes whatever she likes and recomputes the parity row and column to match. Nothing in the scheme stops her, because nothing in the scheme is secret.
That is the entire distinction between a checksum and a message authentication code, and it can be stated in one sentence a ten-year-old can hold:
A checksum catches accidents. What makes it catch enemies is a secret.
This is, in miniature, the same lesson the Italian primary-school pathway discussed in the first article delivers for confidentiality: build, attack, discover the weakness, improve. That follow-on loop is absent from the CS Unplugged activity and uncommon in the materials surveyed here. The card trick is taught as a completed magic trick rather than as a first attempt to be broken, and the natural second act, playing it again against a classmate who knows the rule, is not in the activity.
It should be. It is the cheapest possible introduction to keyed integrity, and it requires no arithmetic beyond odd and even.
Where integrity does get taught, it arrives by side door
Above primary level, integrity does show up, but the pattern is revealing.
CryptoEL, an experiential-learning tool for K–12 cryptography, opens with a hashing module: a high-school student submits a university application, and learners work through what happens when the message is tampered with in transit, arriving at the idea of sending both the message and its hash. It is a well-chosen scenario. The authors are also explicit that digital signatures, HMAC and TLS are left to future work, which is a fair scoping decision and also a precise map of where the field stops.
CryptoScratch goes further, giving learners Scratch blocks for AES, RSA and SHA-256 and setting a task that requires them to create and append a digital signature. Signatures do exist in K–12 material, then. They exist at the very top of it, as a composition exercise for students who have already met symmetric encryption, asymmetric encryption and hashing separately.
And then there is the odd case. Blockchain@School, reported at CHI 2026, teaches data integrity to primary and middle-school learners aged 9 to 13 using physical blocks and cards plus a light web application for validation, including a timed "hacker attack" mode. The authors report workshops with more than 300 learners and evaluate both learning and engagement.
But notice the route. Integrity, one of the four fundamental properties, older than computers and present in seals and signatures throughout recorded history, reaches nine-year-olds because blockchain became interesting enough to build a school toolkit around. The property arrived as a passenger.
There is nothing wrong with that toolkit. There is a gap in a field where, among the young-learner materials reviewed here, unplugged integrity appears either as a magic trick with the wrong attacker or as part of a distributed-ledger activity.
Authentication is taught as conduct, not as mechanism
Authentication is the property children encounter most often and understand least, and here the material is not scarce at all. It is simply about something else.
As the first article noted, primary-school security education centres on passwords: a 2025 survey of 114 Scottish primary teachers found passwords to be the core of what gets taught, with phishing and hacking reserved mostly for older groups. Password education is real, well-intentioned and reasonably widespread.
It is also almost entirely hygiene. Make it long. Use three words. Do not reuse it. Do not tell your friends. Do not write it on the desk.
Every one of those rules is correct. None of them answers the question a child would actually ask if given the chance: what does the computer do with it? Does the school keep a list of everyone's passwords somewhere? If it does not, how does it know mine is right? If somebody steals the list, do they have my password, or something else? Why can the site email me a reset link but not my old password?
A child who has never been given a model of what a password is, a claim checked against a stored derivative of a secret in order to authorise an action, is being asked to follow rules whose reasons are invisible. The Scottish study reports forgetting as pervasive across age groups. It does not establish that the absence of a model causes the problem, but it does show the limits of treating password use only as a set of rules.
The exception, and it is a striking one, comes from the age-banded password work also cited in the first article, which renders the consequence of a leaked password for a six- or seven-year-old as "someone telling the computer that they are you." That is not hygiene. That is a mechanism, expressed at the right level of abstraction, and it is one of the clearest sentences about authentication in the material for young children.
There should be a hundred more like it.
Children are becoming credential subjects faster than they are becoming credential learners
The fourth row of the table, evidence and accountability, is where the timing problem becomes acute. This is no longer a hypothetical part of children's digital lives. It is becoming policy and infrastructure.
California's AB 1043 was signed in October 2025 and becomes operative in January 2027. It requires an operating-system provider to collect an age or birth date at account setup, classify a user into an age bracket, and return a minimal signal to applications through a real-time API. At the federal level, the Digital Age Assurance Act of 2026, S.5090, was introduced on 22 July 2026 and remains a proposal referred to the Senate Committee on Commerce, Science, and Transportation. It would similarly make the operating system the primary source of a user's age bracket. Where technically feasible, its introduced text calls for an interoperable verifiable credential or zero-knowledge proof and defines both in terms of minimal disclosure. In the European Union, the Commission's feature-ready age-verification blueprint lets a user prove that they are over a threshold without revealing identity, birth date or other personal information. It can operate as a stand-alone "mini wallet" and is designed for compatibility with European Digital Identity Wallets.
These initiatives are different in legal status and technical detail. Together, they point toward a recurring interaction in which a device, wallet or other holder presents a bounded claim, and an application or website decides whether to accept it. Some systems use an operating-system signal; others use a signed attestation; some designs support selective disclosure or a zero-knowledge proof. The common point is that children will increasingly be subjects of age-assurance decisions even if they never see the trust machinery behind them.
We conducted a targeted search for educational material that would let a child understand that machinery. The material we found on verifiable credentials in education was about administration: issuing diplomas and micro-credentials, storing badges in a learner's wallet, reducing credential fraud, and letting employers verify a transcript without contacting the registrar. Those are useful applications. They do not address the child as someone who should understand what happens when a credential or age signal is presented on their behalf.
So the position is this. The confidentiality question has had decades of pedagogy and arrives in classrooms as a paper cipher. The credential question is arriving through policy and infrastructure, with little corresponding pedagogy.
From my work with national identity, mobile credentials and digital-certification systems, the recurring mistakes are usually not about algorithms. They are about roles. An issuer makes a claim and signs it. A holder presents it. A verifier decides whether that claim is acceptable for a particular purpose. The identifier encoded in a credential is not the credential, and a successful presentation is not proof that the credential remains valid.
Revocation is where the clean classroom triangle meets operational reality. A signature can remain mathematically correct after the right to use a credential has ended, the document has been replaced or a key has been compromised. The verifier therefore needs current status information, and the design has to decide how that information travels, how old it may be and what happens offline. The hard part is not checking the signature. It is deciding whether the evidence is current enough for the decision being made.
Selective disclosure exposes a different failure. Verifiers routinely ask for the whole credential when they need only one fact. An age check may need "over 18," not a birth date, address and document number. A classroom model must make data minimisation visible: the verifier states the question, the holder releases only the necessary claim, and the class can challenge any request that exceeds the purpose. That is the same distinction children already understand when a hall pass proves permission without announcing why they left.
What the other three properties look like as activities
The constructive claim of this article is that these three properties do not necessarily require more mathematics than confidentiality. They can be modelled with roles, objects and consequences. The following sketches are offered as a research agenda rather than a tested curriculum:
Integrity, with a real adversary. Play the parity trick twice. The first time, somebody flips a card while you look away, and you find it, and it is magic. The second time, tell the class the rule first, then let a volunteer change whatever she likes. She will defeat it, either by the rectangle or simply by recomputing the parity cards. Now add something only you know, a secret subset of cells that count toward the parity, written on a card in your pocket, and let her try again. The lesson can land without the word MAC ever being used.
Detecting is not preventing. A sealed box that anyone can open. The seal stops nobody; it tells you afterwards. Ask which children would rather have: a box nobody can open, or a box that always tells you when it was opened. There is no right answer, and the discussion is the point, because most of the security in their lives is the second kind.
Authentication as a graded set of instruments. One child has to convince another that she is who she says she is, using, in turn: her name; a secret only the two of them agreed; a card the teacher issued this morning; a note the teacher will confirm on request. Then ask, for each one, what happens if it is overheard, copied, lent to a friend, or lost. Four instruments, four failure modes, and the categories something you know, something you have and something somebody vouches for fall out of the discussion rather than being announced.
Issuer, holder, verifier. Three children, one card. The teacher issues a card that says the bearer may borrow from the class library. One child holds it. Another child, playing librarian, has to decide whether to accept it. Ask the librarian how she knows the teacher really issued it, and what she would do if the card were handwritten, and whether she needs to know the holder's name at all to do her job.
That last question introduces selective disclosure through a familiar object: a hall pass can prove permission without revealing every fact about its holder. The trust triangle used in mobile credential systems can be represented as a game with three roles and one piece of card.
Revocation, which is where it gets interesting. The card is lost. Somebody else has it. Now the librarian's problem is not verifying the card, which is fine; it is finding out that the card is no longer good. Children may propose a list of cancelled cards, and then discover that the list has to reach the librarian, be current, and account for a librarian who has not seen today's update. That models the operational core of revocation without requiring cryptographic mathematics.
What not to teach
Continuing the discipline from the first article, a few shortcuts to refuse before they set:
A hash is not encryption. It cannot be reversed, decrypted or "unhashed," and describing it as scrambling invites exactly that misconception. It is a fingerprint of a thing, not a hidden version of it.
A checksum is not a signature. Anyone can compute a checksum. That is the whole difference, and it is the same difference as between the parity trick with a public rule and the parity trick with a secret one.
A signature does not prove when. It proves who, over what. Ordering in time is a separate problem needing separate machinery, and conflating the two produces adults who think a signed document is a timestamped one.
Presented is not valid. A credential that arrives is not a credential that checks out. Somebody has to verify the issuer, and somebody has to know whether it was revoked. "It showed me a green tick" is not verification.
"Verified" is not a property of a person. It is a property of a specific claim, checked by a specific party, at a specific moment, for a specific purpose. Children who learn otherwise become adults who think an account badge means something about a human being.
Encryption does not deliver integrity by itself. A message can be unreadable and still be alterable. This one is genuinely hard, and probably belongs to secondary school, but it should not be actively contradicted earlier.
Where this goes next
The first article argued that cryptography can be taught as a way of reasoning before it is taught as mathematics. This one narrows the claim. Integrity, authentication and credentials can be modelled with cards, boxes, seals and roles. Keyed verification, minimal disclosure, revocation and the separation of issuer from verifier all have plausible physical analogues that can be tested with young learners.
The obstacle is not only the mathematics. The secret message is such a good hook that the field has invested far less in the other three.
The next article in this series goes down to ages six and seven specifically, where the material thins out almost entirely, and where these four properties have to be carried by story, character and consequence rather than by activity design.
References
- CS Unplugged. Error detection and correction: parity magic. See also the classic activity.
- Bell, T., Witten, I. H., & Fellows, M. Card flip magic: error detection and correction.
- Lodi, M., Carrisi, M. C., & Martini, S. (2024). Big Ideas of Cryptography in Primary School. Proceedings of ITiCSE 2024, 206–212.
- Rayavaram, P., et al. (2024). CryptoEL: A Novel Experiential Learning Tool for Enhancing K–12 Cryptography Education. arXiv:2411.02143.
- Percival, N., Rayavaram, P., Narain, S., & Lee, C. S. (2022). CryptoScratch: Developing and Evaluating a Block-based Programming Tool for Teaching K–12 Cryptography Education using Scratch. IEEE EDUCON 2022.
- Rayavaram, P., et al. (2024). Visual CryptoED: A Role-Playing and Visualization Tool for K–12 Cryptography Education. Proceedings of SIGCSE 2024.
- Pellegrino, M. A., & Guasti, L. (2026). Introducing Blockchains at School Without Computers: Hands-On Sense-Making for Young Learners. Proceedings of CHI 2026.
- Lamond, M., Prior, S., Renaud, K., & Wood, L. A. (2025). Teachers' perspectives and practice of cybersecurity education in primary schools. Discover Education, 4, 312.
- Prior, S., & Renaud, K. (2020). Age-appropriate password “best practice” ontologies for early educators and parents. International Journal of Child-Computer Interaction, 23, 100169.
- California Legislature. (2025). AB 1043: Age verification signals: software applications and online services.
- United States Senate. (2026). S.5090: Digital Age Assurance Act of 2026, introduced text.
- European Commission. (2026). The EU approach to age verification.
- European Commission. (2026). Recommendation (EU) 2026/1035 on establishing a common framework for EU-wide age verification technologies.
- Sedlmeir, J., Smethurst, R., Rieger, A., & Fridgen, G. (2021). Digital Identities and Verifiable Credentials. Business & Information Systems Engineering.
