DNA Cryptography: A new Franco-Japanese approach proves its worth

shape

Using DNA as a method to encrypt sensitive messages is now possible. A multidisciplinary team has developed a DNA-based encryption approach that allows for the generation and sharing of random keys to encode messages, regardless of the distance between the sender and the recipient. The method was recently tested for the first time under real-world conditions during the French President’s visit to Japan [1] on April 1, 2026. This work was carried out as part of a collaboration between the CNRS [2], the University of Tokyo, the University of Limoges, IMT Atlantique, and the École supérieure de physique et de chimie industrielles de la ville de Paris (ESPCI Paris – PSL), with support from the ANR and France 2030 [3]. The findings have been pre-published in an open-access repository [4].

Protecting confidential communications: A major challenge in the digital age

Today, the encryption of sensitive data relies primarily on so-called “conditional” methods, whose security is based on the assumption that no external actor has sufficient computing power to break the code. However, other so-called “unconditional” approaches exist, such as Vernam encryption (or the OTP—“One-Time Pad”—method) [5]. Although it offers perfect security—in the sense that it guarantees that security does not depend on an adversary’s computational power—this approach imposes several constraints: the key used to encrypt the message must be shared in advance between the sender and the recipient. It must also be as long as the message itself, used only once, and “perfectly” random—that is, impossible to predict. However, generating and sharing large, one-time-use random keys remains very difficult using existing methods, particularly when the sender and recipient are separated by significant distances.

DNA for encrypting messages

This is where DNA gets interesting. Each DNA molecule is composed of four chemical bases (A, T, C, and G), and chemists are able to commercially synthesize long chains in which the order of the bases is statistically random. These DNA sequences can then be copied identically using enzymatic processes and thus shared between a sender and a recipient.

Specifically, scientists prepare sets of duplicated DNA strands—entirely synthetic in origin—[6], one copy of which is kept by the sender and the other by the recipient. The DNA fragments they contain will allow the correspondents to generate perfectly random encryption keys, which will nevertheless be identical in pairs. This is done just before communication begins, using powerful sequencing machines that read the molecules to assemble a binary digital key (composed of 0s and 1s) that can be used to encode, send, and decode a message up to several hundred megabytes in size.

A reliable, secure, and high-performance method, even over long distances

What are the strengths of this approach? DNA offers remarkable storage density and stability: when properly preserved, the polymer can remain intact for thousands of years, and just a few milligrams are enough to store exabytes of binary information—the equivalent of a million hard drives. Furthermore, generating shared cryptographic keys using DNA has the advantage of being independent of the distance between the sender and the receiver. In other words, there would be nothing preventing this method from being used between Earth and the Moon, or even beyond.

But above all, this DNA-based approach makes the only cryptographic method that can be mathematically proven to offer unconditional security—that is, security independent of an adversary’s computational power—more accessible. By testing various scenarios, scientists have demonstrated that even if the DNA used to generate the keys were intercepted, the channel would remain unbreakable: since there are only two copies of each DNA sequence—one for the sender and one for the recipient—any partially stolen key would never be reused by the correspondents. Similarly, if the eavesdropper attempted to amplify the key to obtain multiple copies before returning it to the users, this manipulation would result in discrepancies in the number of copies, which could be detected by the correspondents, who would then decide to stop using those keys.

Thanks to its various strengths and reliability, this approach opens up new possibilities for protecting the most sensitive communications, whether diplomatic, military, or scientific. In the longer term, it could also find applications in extreme environments, particularly in space communications or critical digital infrastructure, where the reliability and tamper-resistance of communications are major concerns.

Notes

[1La démonstration se fera lors de la visite du Président au « Laboratory for Integrated Micro-Mechatronics Systems » basé à Tokyo.

[2Du laboratoire Gulliver (CNRS / ESPCI PARIS – PSL). D’autres scientifiques du Laboratory for Integrated Micro-Mechatronic Systems (CNRS / Université de Tokyo) sont impliqués, dont l’un des porteurs du projet, Anthony Genot, aujourd’hui décédé (https://www.insis.cnrs.fr/fr/cnrsinfo/hommage-anthony-genot-une-intelligence-en-resonance-avec-lavenir).

[3Dans le cadre du PEPR MoleculArXiv, piloté par le CNRS : https://pepr-molecularxiv.fr/ et du projet ANR DNA Sec, porté par IMT Atlantique.

[4https://hal.science/hal-05560338
Ces travaux n’ont pas encore été validés par une revue scientifique à comité de lecture.

[5Le chiffrement de Vernam, ou méthode « One-time pad » (OTP), est un système de chiffrement symétrique utilisant une clé aléatoire de même longueur que le message, utilisée une seule fois. Lorsqu’il est appliqué correctement, il offre une sécurité théoriquement parfaite, car le message chiffré ne révèle aucune information sans la clé.

[6L’ADN utilisé en cryptographie est issu d’un processus de fabrication synthétique qui s’inspire uniquement du principe de codage de l’ADN, sans aucun lien biologique, fonctionnel ou génétique avec l’ADN des organismes vivants.

References

Synchronized DNA sources for unconditionally secure cryptography. Sandra Jaudou*, Hélène Gasnier*, Elias
Boudjella*, Marc Canève, Victoria Bloquert, Vasily Shenshin, Tilio Pilet, Sacha Gaucher, Soo Hyeon Kim, Philippe Gaborit,
Gouenou Coatrieux, Matthieu Labousse, Anthony Genot, and Yannick Rondelez. *contributions équivalentes
Lien HAL : https://hal.science/hal-05560338

Key information

icon Published on 31/03/2026

icon Research

icon Chercheur CNRS | Yannick Rondelez | yannick.rondelez@espci.fr / Chercheur CNRS | Matthieu Labousse | matthieu.labousse@espci.fr

icon Presse CNRS | Elisa Doré | T +33 1 44 96 53 16 | elisa.dore@cnrs.fr