By Luigi Caramico, Founder and CTO of DataKrypto

As we prepare to kick off Cybersecurity Awareness Month, I want to shine the spotlight on a historic figure to whom much can be owed when it comes to modern-day cryptography: Leon Battista Alberti.

Those who run consequential systems understand how rare genuine trust is. In the 15th century, the Genoese polymath Leon Battista Alberti observed that leaders wished for a confidant worthy of knowing their deepest plans — but, “because of the common treachery of men,” such a person was hard to find. The answer, he argued, was not better people but better methods of writing — what he called ciphers.

Six centuries later, in an era of AI agents and borderless cloud systems, his observation lands with fresh urgency. Because we cannot change human nature, we engineer around it. Cryptography is the way we create and bottle trust and move it across untrusted networks, devices, and now – AI models.

The Evolution of Cryptography: From Alberti to Today

In 1466, Alberti wrote De Cifris, one of the first systematic works on cryptography. He argued that only ciphers could safeguard secrets, because people could not.

Alberti was more than a cryptographer. He was a true Renaissance figure: an architect, painter, and mathematician. His invention of the polyalphabetic cipher – an encryption technique that uses multiple substitution alphabets to encrypt a message – and the cipher wheel – used to scramble a message (encryption) or unscramble it back into its original form (decryption) – transformed secrecy from a fragile practice into a reliable method. For Alberti, cryptography was not just a technique. It was a way to engineer trust when trust in people was impossible.

Alberti’s work set off a chain of breakthroughs that still influence cybersecurity today:

• 1550: Girolamo Cardano invented the grille – a template, typically a card with cut-out holes, that is placed over a sheet of paper to write or read a plaintext message through the apertures
• 1553: Giovan Battista Bellaso published a polyalphabetic cipher that used a repeating keyword to mix cipher alphabets. This is the cipher now known as the Vigenère cipher.
• 1790s: Thomas Jefferson advanced cipher wheels,
• 1920s: Arthur Scherbius invented The Enigma machine* – an electro-mechanical rotor cipher machine later used by Nazi Germany during World War II to encrypt military communications
• 1930s/1940s: Bletchley Park in operation* – the former top-secret home of the World War Two Codebreakers, including many famous Codebreakers such as Alan Turing, Gordon Welchman, and Bill Tutte
• 1945: Claude Shannon defined secrecy mathematically
• 1976: Whitfield Diffie and Martin Hellman created the Diffie-Hellman key exchange
• 1977: Ron Rivest, Adi Shamir and Leonard Adleman invented RSA encryption
• Present: TLS and AES invisibly secure the internet, while zero-knowledge proofs, multiparty computation, and post-quantum cryptography have moved from theory into practice

* These innovations proved the influence of math and machinery on the outcome of wars. Today, the pattern is constant: fragile trust in people can be replaced with verifiable guarantees established through mathematics.

Cryptography in the AI Era

That lesson still applies. Today, the confidants of our secrets are not only humans but also AI models and agents. They process sensitive data, generate outputs, and make inferences at scales Alberti could never have imagined. Like courtiers of old, they cannot always be trusted.

If we cannot change their nature, we must constrain their behavior with cryptography:

• Keys are sovereign. Protect them with hardware roots, minimize exposure, and rotate often.
• Proof beats trust. Verify everything: code, data, runtime.
• Divide powers. Split roles and enforce policy with cryptography, not goodwill.
• Minimize exposure. Use just-in-time access, scoped tokens, zero-knowledge proofs, and privacy-preserving analytics.
• Assume your model is curious. Treat prompts and outputs as sensitive, even if nothing is stored.
• Stay agile. Cryptography evolves. Systems must be designed so primitives can be swapped, especially as post-quantum standards arrive.

Building Trust with Math

As an Italian cryptographer, I take pride in Leon Battista Alberti as part of my country’s heritage and our profession’s history. His genius was to recognize a permanent truth: people are fallible, but systems can be made reliable.

Six centuries on, his warning still resonates. When the trustworthy person is rare, the trustworthy method must be standard. That is the role of cryptography in the AI era: to turn trust from an act of faith into a mathematical certainty.