Two Cryptographers Built a Coded Message That Physically Cannot Be Copied — Split It in Two and Both Halves Become Coin Flips

Classical encryption has a structural weakness that has nothing to do with how good the math is: ciphertext is data, and data can be copied. An adversary who intercepts an encrypted message can keep it indefinitely, duplicate it freely, and wait for the key or for a computer capable of breaking the scheme. Everything encrypted today is exposed to whatever cryptanalysis exists in twenty years.

Prabhanjan Ananth of the University of California, Santa Barbara and Amit Sahai of UCLA have described a scheme that removes that option. Their approach, posted as a preprint on arXiv, encodes the message not as bits but as quantum states — and quantum states obey the no-cloning theorem, the result that an unknown quantum state cannot be duplicated. There is no copy to keep.

The construction is deliberately minimal. A single-bit message, a 0 or a 1, is spread across multiple quantum states using a secret digital key that determines how the information is distributed among them. A receiver holding the matching key measures all the states together and recovers the bit. The information is not stored in any individual state; it exists only in the relationship among them.

That is what makes the interesting attack fail. Suppose an eavesdropper intercepts the transmission and, unable to copy it, splits the states between two accomplices so that each holds part of the ciphertext. Each accomplice must eventually measure what they hold, and measurement is destructive — it collapses the state and irreversibly disturbs the correlations that carried the message. When the key later becomes public, neither accomplice can reconstruct the bit. Their success rate converges on 50/50, indistinguishable from guessing.

The authors state that "the scheme is exponentially secure," meaning the adversary's advantage shrinks exponentially as more quantum states are folded into the encoding. Security here does not rest on an assumption that some mathematical problem is hard; it rests on the measurement postulate of quantum mechanics. A more powerful computer does not help, because computation is not the bottleneck.

Earlier attempts at uncloneable cryptography ran into efficiency problems severe enough to keep them theoretical — schemes that worked on paper but demanded impractical resources or unrealistic assumptions about the quantum channel. This work addresses those limitations, which is the reason it drew attention on the day it was posted.

Practical deployment remains distant. Encoding a single bit across many quantum states, transmitting them without decoherence, and measuring them jointly at the far end is an engineering problem that current hardware handles only over short distances and in laboratory conditions. But the theoretical result is the harder half, and it establishes something classical cryptography can never offer: a message that an adversary cannot save for later.