Anamorphic Cryptography via Randomness Substitution

26 concepts13 papers1 proposal

Imagine you are forced to communicate under the watchful eye of a powerful adversary—a "dictator" who monitors all network traffic, bans unapproved encryption, and can seize your private cryptographic keys at any moment. If you send an obviously encrypted message, you arouse suspicion. If you are forced to use a dictated message or standard protocol, your capacity for private communication is stripped away.

How do you hide a secret message in plain sight when even your underlying keys are not safe?

The answer lies in a powerful realization: modern cryptographic primitives are factories of pure noise. Whether it is the internal randomness used to blind an RSA ciphertext or the unique nonce generated for an elliptic curve digital signature, cryptographic outputs are designed to look like high-entropy, statistically random bitstreams.

Anamorphic cryptography and subliminal signaling exploit this exact property. By systematically replacing this expected "cryptographic noise" with a pseudorandomly encrypted secret message, users can establish completely undetectable, deniable communication channels. Because an outside observer expects nothing but random bits, the altered ciphertext or signature remains mathematically valid and statistically indistinguishable from standard traffic.


Subliminal Channels in Digital Signatures

To understand how this works in practice, consider standard digital signature algorithms like the Elliptic Curve Digital Signature Algorithm (ECDSA) or Schnorr signatures, which underpin the vast majority of modern blockchain transactions and authentication protocols.

To sign a message, these algorithms require the signer to generate a fresh, completely random secret value called a nonce. If an observer looks at a valid signature on a public ledger, the nonce field appears as a string of maximum-entropy data.

Standard Signature:  [ Message ] + [ High-Entropy Random Nonce ]
Covert Signature:    [ Message ] + [ Pseudorandom Ciphertext (Hidden Message) ]

As demonstrated by Tiemann et al. and Chen et al., a sender can intentionally replace this random nonce with a pseudorandom ciphertext containing a hidden message. Because a properly encrypted ciphertext is statistically indistinguishable from true random noise, the resulting signature passes all standard verification checks and statistical analyses.

This capability is not a rare flaw; rather, it is an inherent byproduct of common cryptographic design patterns. Persiano et al. showed that standardized signature protocols derived from the Fiat-Shamir transform (such as RSA-PSS) naturally support these hidden channels. Crucially, the authors proved that even if a dictator forces you to disclose your primary signing keys, the presence of the hidden data cannot be proven. The embedded ciphertext still looks exactly like the legitimate cryptographic noise that would have been generated by a standard, uncompromised signing process.


Anamorphic Encryption and Dual Decryption

While signatures allow you to hide data inside authentication traffic, anamorphic encryption applies this concept directly to public-key encryption. First formalized by Phan et al., anamorphic encryption creates a "dual decryption" pipeline out of a single ciphertext.

If a dictator seizes your standard private key, the ciphertext decrypts cleanly into a completely benign, dictated message. However, if the ciphertext is processed using a separate, secret "double key," it reveals an entirely different, hidden message.

                  ┌───> Standard Key ───> Benign Message (For the Dictator)
                  │
[ Ciphertext ] ───┤
                  │
                  └───> Double Key   ───> Hidden Secret Message

Why is this dual-decryption capability so prevalent? Persiano et al. proved that anamorphism is a latent feature in many of the world's most common encryption schemes, including RSA-OAEP and ElGamal.

The mechanism works by exploiting internal randomness recovery. In a scheme like RSA-OAEP, the decryption process naturally allows the receiver to reverse the cryptographic math and extract the exact random bits used by the sender to blind the plaintext. Because the receiver can recover this internal randomness, the sender can replace those random bits with an encrypted payload. To the dictator, the ciphertext looks perfectly normal; to the true recipient, the recovered "randomness" is actually the second ciphertext, waiting to be decrypted by the double key.

This exploitation of mathematical redundancy takes many forms. For instance, Chang demonstrated that the existence of multiple valid mathematical roots for a single ciphertext provides a natural, zero-overhead channel for embedding hidden metadata without altering file sizes or compromising security proofs.


Eliminating Pre-Shared Secrets

Early iterations of these covert channels suffered from a major logistical flaw: the sender and receiver had to securely exchange a secret key beforehand to coordinate how the randomness would be perturbed or decrypted. If you are already under total surveillance, exchanging a pre-shared key is nearly impossible.

To solve this, Phan et al. introduced public-key anamorphic encryption. Under this framework, a sender can embed a hidden message using only the recipient's standard, publicly available encryption key. No prior coordination or shared secrets are required. The recipient, upon receiving the ciphertext, uses a private secondary key to extract the hidden payload.

The authors proved that this can be achieved even in highly secure, chosen-ciphertext attack (CCA) secure systems—such as the Koppula-Waters encryption scheme—without altering their underlying security architectures.

To achieve this generically across hybrid encryption systems, Wang et al. proposed a method that perturbs the randomness within Key Encapsulation Mechanisms (KEMs). By leveraging the homomorphic properties of the underlying encapsulation math, the sender can inject secret bits into the key exchange portion of a transmission. A secondary recipient with the appropriate key can intercept and extract the hidden bits, while the primary ciphertext remains fully valid and indistinguishable from standard network traffic.


The Fundamental Law of Covert Signaling

If cryptographic noise can be substituted so easily, what prevents users from hiding massive files inside every packet?

Covert channels are strictly bound by the statistical limits of detection. Soltani et al. established the Square Root Law for covert communication, proving mathematically that a sender can safely hide at most $O(\sqrt{n})$ bits of information in a stream of $n$ packets without giving an adversary enough statistical evidence to confirm the channel's existence.

The authors emphasized that signaling mechanisms remain secure only as long as they perfectly mimic the natural entropy of their environment. If you alter a field—such as the least significant bits of a high-resolution timer or a packet size parameter—the alteration must look exactly like the random variance inherent to that system.

This core rule scales across every domain of information hiding:

  • Algebraic Surjectivity: Chen et al. developed public-key steganography frameworks that utilize the entire set of points on an elliptic curve. By ensuring total coverage (surjectivity) rather than using a subset of points, the hidden data maps perfectly to the expected distribution of random curve points, allowing the system to support advanced primitives like identity-based encryption covertly.
  • Protocol Compatibility: Liu et al. demonstrated that authentication data can be "drip-fed" through the existing data fields of legacy aviation transponder protocols (ADS-B). By mimicking standard protocol noise, upgraded systems can verify identities while remaining completely transparent to older hardware.
  • Textual Hash Collisions: Even outside pure cryptography, Wang et al. showed that command-and-control messages can be embedded into benign social media text by exploiting hash collisions, breaking down sensitive data like IP addresses into pieces that masquerade as ordinary textual variations.

Ultimately, anamorphic cryptography works because it turns an adversary's demands against them. By forcing users to adopt strong, high-entropy cryptographic standards to sign or encrypt their data, a warden inadvertently creates the perfect, high-entropy camouflage needed to hide private conversations in plain sight.

Key Foundational Work

Applying Information Hiding in VANETs to Covertly Report Misbehaving Vehicles (2014) — contributed 2 concept elements to this direction.
Fundamental Limits of Covert Bit Insertion in Packets (2018) — contributed 1 concept elements to this direction.
History & citationVersion 2 · 2026-10-04

Current · Permanent link

  1. Version 2 (latest) · 2026-10-04 · 11 papers
  2. Version 1 · 2026-10-04 · 10 papers
Cite this direction

Latest version (a living page):

@misc{lacuna_anamorphic_encryption_6e3dba69,
  title = {anamorphic encryption},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/anamorphic-encryption/rdr_6e3dba69e0454e9ca85a6e62e8046806},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}

This exact version (version 2, frozen text):

@misc{lacuna_anamorphic_encryption_6e3dba69_v2,
  title = {anamorphic encryption},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/anamorphic-encryption/rdr_6e3dba69e0454e9ca85a6e62e8046806/v/2},
  note = {Version 2, frozen 2026-10-04. Accessed 2026-11-20},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake