CRYPTOGRAPHY EDUCATION · PART 1

Children are already being taught to hide messages. They are not being taught how systems decide what to trust.

Cryptography education, part 1. This article maps how cryptographic ideas enter primary and secondary education. Later parts will narrow the analysis to specific age groups and trust concepts.

The question that is already answered

Cryptography is usually introduced late. In formal education it tends to appear after programming, discrete mathematics, modular arithmetic or computer networks, on the reasonable assumption that modern cryptography is mathematically demanding and students need a technical foundation first.

That assumption confuses two different things. Understanding a cryptographic algorithm is difficult. Understanding the problem cryptography is trying to solve is not. A child does not need modular exponentiation to grasp that a message should sometimes be readable by one person and not another, and a student does not need to understand a hash construction to recognise that changing a message after it was sent is a different act from merely reading it.

Over the last two decades, educators in several countries have tested that distinction. Published interventions show that primary and secondary students can encrypt, decrypt and attack messages. They can reason about statistical patterns in ciphertext. They can model public-key systems without implementing one. They can do it through paper puzzles, card games, block-based programming, escape rooms and national competitions. On the narrow question of whether children can engage meaningfully with cryptographic ideas, the evidence supports a clear answer: yes.

So the interesting question is no longer whether. It is what, and the examples surveyed here cluster around a surprisingly narrow slice of the subject.

Most of the reviewed material enters through secrecy. Far less enters through integrity, authentication, identity, credentials or trust. That gap is the subject of this article.

Everything enters through the secret message

Most introductions to cryptography for children begin with the same hook, and for a good reason. A substitution cipher converts an abstract security problem into an observable transformation. A student writes something, applies a rule, obtains something unreadable, hands it to somebody else and watches the process run in reverse. The entire system fits on a piece of paper.

This approach appears prominently in CryptoClub, developed originally at the University of Illinois Chicago and now based at the University of Chicago, and supported through multiple U.S. National Science Foundation grants. CryptoClub uses secret messages to connect cryptography with middle-school mathematics: encrypting and decrypting, breaking ciphers, relay races, treasure hunts and digital games. Its materials were tested iteratively over six years, including a 2012–2013 field test at 29 sites across the United States.

The attraction of the secret message is real. But the educational value comes from what happens next, when somebody asks: can you break it?

That question changes the activity. A cipher stops being a transformation rule and becomes a system with an adversary. Patterns matter. Repetition matters. Language statistics matter. Key length matters. Students begin to discover that making something look unintelligible is not the same as making it secure. This is the first genuine conceptual transition in the subject, and it is the one that separates a decoder-ring activity from cryptographic thinking.

The strongest existing design is adversarial

The clearest example of that design comes from Italy. In 2024, Michael Lodi, Maria Cristina Carrisi and Simone Martini reported a learning path built and tested with fifth-grade primary students. Rather than presenting techniques as a sequence of historical curiosities, the activities were arranged so that limitations discovered in one system motivated the introduction of the next, combining unplugged exercises with task-specific block-based tools. The published materials include Caesar encryption, frequency analysis against Caesar, and an interactive simulation of a block cipher. The same group had previously run a comparable intervention at Grade 10 using Snap! and unplugged activities, including a remote-unplugged Diffie–Hellman.

The important part is not Caesar. It is the sequence:

build → attack → discover the weakness → improve.

A student first believes that replacing each letter with another letter protects a message. Then an attacker finds statistical structure and the protection fails. A stronger construction has to conceal or eliminate that structure. The failure is not a detour in the lesson. The failure is the lesson, and it is the same intellectual move that modern cryptography is built on: define what an adversary can do, decide which property must survive that adversary, construct a mechanism that satisfies it. The mathematics changes enormously with age. The structure does not.

Cryptography also happens to supply every ingredient of a game: a secret, a rule, incomplete information, an opponent, evidence and a goal. That is why the same design keeps reappearing in narrative form. In Ireland, researchers at Technological University Dublin built The Code of Kells, a collaborative game for primary students aged roughly 10 to 12 who investigate the theft of the Book of Kells across a map of Dublin, recovering clues protected by Caesar, Polybius and Pigpen ciphers; eighty children took part in the evaluation.

France runs the same idea at national scale. The Concours Alkindi, organised by Animath and France-ioi with support from the Ministère de l'Éducation nationale, Inria and the DGSE, introduces pupils in France's 4e, 3e and 2nde classes to cryptanalysis through progressively harder rounds. The tenth edition, in 2024–2025, drew a little over 52,000 students; the qualifying rounds asked them to find codes hidden in images, recover an encrypted password and work through stream-cipher exercises, and the final brought together 20 teams and 74 qualified students, 47% of them young women.

By this point the educational object has moved. Students are not being told what encryption is. They are being placed on the other side of it.

The abstraction can be separated from the implementation

The separation between a cryptographic idea and its production implementation is clearest in public-key cryptography. The classic CS Unplugged activity Kid Krypto models public-key encryption without RSA, elliptic curves or any real cryptosystem. Students publish information that can be used to lock a message while keeping private the information needed to unlock it. The activity is aimed at junior-high level and above and can be run by a whole classroom, including students acting as interceptors.

The implementation has been abstracted away. What survives is the property:

information used to protect a message does not have to be secret.

For a student whose mental model is padlocks and shared passwords, that is a surprising claim, and the surprise is the point. It delivers one of the central breaks between classical and modern cryptography without any of the mathematics that makes a real system secure.

Curricula are moving from safe use toward understanding

In the programmes and standards reviewed here, cryptography does not appear as a standalone school subject. It sits inside computer science, mathematics, digital literacy or cybersecurity. But the direction of travel is visible in the standards themselves.

The 2026 CSTA PK–12 Computer Science Standards, released in July 2026, merged the previous Computing Systems and Networks & the Internet concepts into a single Systems & Security concept, explicitly to reflect their interconnection and the growing importance of security in responsible use of computing. Cybersecurity is also defined as one of the high-school specialty pathways, with sub-areas including threats and security measures, cybersecurity policies, and network theory and design. The earlier 2017 standards were already explicit at middle-school level, where students in grades 6–8 were expected to model several methods of encryption, from substitution and steganography to unplugged models of public-key encryption.

Code.org takes older students further, asking them to experiment with symmetric and asymmetric encryption and to consider data at rest, in transit and in use, including the trade-offs among security, performance and usability. New Zealand's digital technologies curriculum places encryption among the computer science concepts students analyse in authentic contexts, alongside error control and computational complexity. The UK national curriculum is less prescriptive about mechanisms but requires secondary students to understand secure use of technology, online identity and privacy.

These are different educational models, but the transition is the same one. Security is becoming something students are expected not merely to obey, but to understand.

Encryption is only one of four questions

There is a risk in teaching all of this through secret messages. Students can leave with the impression that cryptography equals encryption equals hiding text. Modern systems do considerably more.

Question a child can ask	Cryptographic 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

Digital signatures can contribute to that fourth property, but non-repudiation is not supplied by a signature primitive alone. It depends on the surrounding identity, key-management, procedural and legal context.

A Caesar cipher gives children an intuitive entry into the first question and says almost nothing about the other three. Meanwhile the digital life they already inhabit is dominated by the other three. A student logs into an account. A device asks for a password. A phone recognises a face. A browser decides whether to trust a site. A messaging app claims a conversation is encrypted. A game server decides whether a request is legitimate. An application update has to be distinguished from a modified one.

Every one of those is an authentication, integrity or trust decision, and children are making them daily, years before they have words for the properties involved.

Uruguay illustrates the shape of the gap

The pattern is easy to see close to home. Ceibal's Ciudadanía Digital classroom materials for 5th and 6th grade of primary education cover digital identity and footprint, privacy, security and online relationships. Its Pensamiento Computacional e Inteligencia Artificial programme runs with 4th, 5th and 6th grade groups, and there is a separate Ciencias de la Computación strand.

So identity, privacy and security do arrive in Uruguayan primary classrooms, and early. They arrive framed as conduct: what you should share, what you should protect, how you should behave. What does not arrive is the mechanism question. A child can be taught to keep a password private without ever being asked what a password does, or how a machine decides that the person holding it is the person it belongs to.

That is not a criticism of digital citizenship, which is doing a different and necessary job. It is an observation about where the vacancy is.

The youngest end is thinner, and thinner in a specific way

Research becomes sparse below about eight years old, though less sparse than it first appears, and the shape of what exists is informative.

A 2025 study by Lamond and colleagues surveyed 114 primary teachers across 14 Scottish local authorities and found that cybersecurity education in primary school centres on passwords, with topics such as phishing and hacking taught predominantly to older groups. Teachers reported that forgetting was a pervasive problem across all ages, and that the support strategies they resorted to may actually reinforce poor password practice.

Separately, work on age-appropriate password education has produced principle sets banded for children aged 4–5, 6–7 and 8–9, with the full set intended to be in place by around nine. One detail from that work is worth dwelling on: rather than using the word impersonation, the material renders the consequence of a leaked password as "someone telling the computer that they are you."

That is a striking sentence to find in material for six-year-olds, because it is not a confidentiality lesson. It is an authentication lesson. The youngest end of the field is thin on cryptography, but where it exists at all, it is already reaching for exactly the properties the older material neglects.

Which suggests that the scarcity of cryptographic material for six- and seven-year-olds is not evidence that the ideas are unsuitable. It is evidence that the usual vehicle is unsuitable. A six-year-old does not need a cipher wheel to encounter confidentiality; two envelopes will do. Nor does a seven-year-old need the word authentication to answer:

• How could she know the message really came from her friend?
• What could we put on this box so we'd notice if somebody opened it?
• Is your name enough to prove that you're you?

Those three questions contain, in order, the intuitions behind authentication, tamper evidence and the distinction between an identifier and a proof of identity. At this age the objective is not to teach primitives early. It is to build the distinctions the primitives will later depend on:

• A secret is different from an identity.
• An identity is different from a proof of identity.
• Reading is different from modifying.
• Encoding is different from protecting.
• Holding a credential is different from knowing a password.

Those are simple enough to appear in a children's story. Their consequences run all the way to the hardest problems in identity engineering.

A progression organised by abstraction, not by algorithm

Existing programmes suggest a possible progression. It should not be mistaken for a curriculum, and the governing variable is not age but the level of abstraction expected of the learner.

Stage	Trust question	Available concepts
Early primary	Who should be able to see this, and how would we know?	private and public information, rules, identity, permission
Primary	How can we hide this message?	transformation, key, encryption and decryption, patterns
Late primary	Could somebody break it, or change it without us noticing?	adversary, cryptanalysis, integrity, randomness
Early secondary	How can two people communicate securely in the open?	symmetric keys, public and private information, threat models
Secondary	How do digital systems establish trust between strangers?	hashes, authentication, signatures, certificates, protocols
Upper secondary	What makes the construction secure?	complexity, entropy, number theory, security definitions

What changes along that progression is the representation, not the subject. A six-year-old manipulates envelopes and cards. A ten-year-old attacks a substitution cipher. A thirteen-year-old simulates public-key encryption. A sixteen-year-old implements a cipher or reasons about key exchange. Eventually a university student studies security definitions, reductions and proofs. There is no need for a discontinuity between those experiences, and the trust question in the middle column is what makes them continuous.

What not to teach matters as much as what to teach

Introducing cryptography early carries an obvious risk: a simplification that produces a misconception which is then expensive to remove. A few shortcuts are worth refusing from the start.

Encoding is not encryption. Turning text into binary, Base64 or Morse changes its representation. It provides no confidentiality whatsoever, and conflating the two is probably the single most common error in introductory material.

A password is not an encryption key. Their roles can overlap in a particular system, but the concepts are not interchangeable, and treating them as such makes key management incomprehensible later.

A biometric characteristic is not a secret. A face or a fingerprint is an identifier or an authentication factor. Unlike a password, it cannot be replaced after disclosure, which makes it categorically different from anything else a student will have met.

Historical ciphers are not contemporary security. Caesar and Vigenère are excellent educational objects precisely because students can break them, and they should be presented as breakable from the outset.

Complexity is not security. A transformation a child cannot reverse is not therefore cryptographically secure. Obscurity is not a property that survives an adversary.

Cryptography is not magic performed by software. The learner should meet both a defender and an attacker. A construction should have a purpose, an adversary should have stated capabilities, and success should mean that a specific property survived a specific attack. That is the whole bridge from a classroom puzzle to security engineering.

The mechanism can be part of the lesson

One more possibility follows from all of this, and it applies to anybody building educational software rather than worksheets. The security concepts do not have to live only inside the exercise. They can live in the interaction model around it.

Consider something as ordinary as entering an educational platform. A child could pick a character, enter a short secret, use a card the teacher handed out, or present a QR credential. The pedagogical question then shifts from "how do you log in?" to "how did the computer decide that you were you?" Once that question is on the table, the mechanisms become comparable because they fail differently. A secret can be guessed. A card can be copied or lent. A username can be known by somebody who is not its owner. A face cannot be reissued.

For young learners those ideas need characters, objects and consequences rather than terminology. The underlying question is authentic regardless.

The question worth asking before algebra

The objective here is not to turn primary students into cryptographers, teach six-year-olds RSA, or reduce modern cryptography to decoder rings. It is to give students progressively better models for answering one question:

When information moves between people and machines, what would have to be true for us to trust what happened?

Most existing material answers a narrower version of that question, "can somebody else read this?", and answers it well. The systems these children will actually inhabit depend at least as much on integrity, authentication, identity, authorisation, randomness and evidence. Those properties receive comparatively little attention in the early-years literature surveyed here, and much of what does exist arrived through password education rather than through cryptography.

That is a gap worth working in, and it is where this line of research goes next.

---

References

1. Lodi, M., Carrisi, M. C., & Martini, S. (2024). Big Ideas of Cryptography in Primary School. Proceedings of ITiCSE 2024, 206–212. Open-access manuscript.
2. Carrisi, M. C., Lodi, M., & Martini, S. Cryptography teaching materials for primary school.
3. Lodi, M., Sbaraglia, M., & Martini, S. (2022). Cryptography in Grade 10: Core Ideas with Snap! and Unplugged. Proceedings of ITiCSE 2022, 456–462.
4. University of Chicago STEM Education. CryptoClub Project.
5. Rocha, M., Almo, A., & Dondio, P. (2022). Who Stole the Book of Kells? Description and Player Evaluation of a Cryptography Game for Primary School Students. Proceedings of the 16th European Conference on Games Based Learning.
6. Bell, T., Witten, I. H., & Fellows, M. CS Unplugged: Kid Krypto, Public-Key Encryption.
7. Concours Alkindi. See also the DGSE account of the 2024–2025 competition.
8. Computer Science Teachers Association. (2026). 2026 CSTA PK–12 Computer Science Standards. About the standards. https://doi.org/10.1145/3820482
9. Code.org. Encryption, AI Foundations: Exploring AI and CS.
10. New Zealand Ministry of Education. Computational Thinking, Progress Outcome 7.
11. UK Department for Education. National Curriculum in England: Computing Programmes of Study.
12. Lamond, M., Prior, S., Renaud, K., & Wood, L. A. (2025). Teachers' perspectives and practice of cybersecurity education in primary schools. Discover Education, 4, Article 312.
13. Prior, S., & Renaud, K. (2020). Age-appropriate password "best practice" ontologies for early educators and parents. International Journal of Child-Computer Interaction, 23–24, 100169. Accepted manuscript.
14. Ceibal. Ciudadanía digital en el aula.
15. Ceibal. Pensamiento Computacional e Inteligencia Artificial. See also Ciencias de la Computación.