How can we trust photos in a world of deepfakes?

In today's digital landscape, photos are no longer a guarantee of truth. With AI, realistic images can be created that never existed. To fight this, scientists have designed Post‑Quantum VerITAS, a cryptographic system that certifies an image's authenticity, even when quantum computers threaten traditional methods. The core of the project is a lattice‑based hash algorithm, resistant to future quantum attacks. When a photo is taken, its creator signs it digitally; any later edit, like cropping or resizing, generates a new proof showing the continuity of its origin chain. These proofs are built with zk‑SNARK, a technique that lets anyone verify validity without revealing underlying data, ensuring privacy and speed. The result is a verifiable certification without third‑party authorities. Journalists, social platforms, and users can check in seconds whether an image comes from a reliable source or has been tampered with. In practice, Post‑Quantum VerITAS turns every photo into a "digital passport" that also resists future quantum computers, making it harder to spread fake news based on deceptive images.

Post‑Quantum VerITAS rests on three cryptographic pillars: lattice‑based hashing, modified Poseidon functions, and zk‑SNARK proofs. Lattice‑Based Hashing leverages hard mathematical problems such as Learning With Errors (LWE), considered infeasible for quantum computers, ensuring the image's digital fingerprint remains immutable. Poseidon, originally designed for zero‑knowledge circuits, has been adapted to operate on larger finite fields, improving GPU computation efficiency and lowering energy consumption. zk‑SNARK (Zero‑Knowledge Succinct Non‑Interactive Argument of Knowledge) proofs allow one to demonstrate that a specific editing operation (cropping, resizing, compression) was performed on a signed image without revealing the image itself or the operation details. This is crucial for privacy, as users need not share original content to obtain verification. The protocol includes a trusted setup phase that generates public parameters and a verification key. Afterwards, the signer creates a digital signature using a post‑quantum key (e.g., NTRU or Dilithium). Each edit generates a new commitment, and the verifier runs a single zk‑SNARK verification in a few milliseconds. Scalability is ensured because proofs have constant size (≈200 bytes) regardless of image dimensions. Moreover, the system works with standard image formats (JPEG, PNG) via a metadata layer that embeds signatures and commitments. In practice, a journalist can upload a photo to a platform that automatically adds a VerITAS signature; readers, using a browser extension, can verify authenticity in real time, receiving a green signal if the trust chain is intact or red if broken. This immediate verification curtails the spread of visual deepfakes, as it does not rely on centralized third parties. Finally, quantum resistance is validated through simulations on small‑scale quantum circuits, showing that even with 50 qubits an attack remains impractical, confirming the model's robustness for the coming decades.