What is AMPC | World ID's Quantum-Safe Security
Anonymous Multi-Party Computation: Protecting Data with Next-Generation Cryptography
1. What is AMPC
AMPC (Anonymous Multi-Party Computation) is the next-generation cryptographic system that powers World ID's security.
In Simple Terms
AMPC distributes your iris data across multiple independent institutions, ensuring that no single entity can access the data alone.
Why AMPC is Needed
🎯 The Problem
In traditional systems, a single organization holds all data. If that organization is hacked or acts maliciously, there's a risk of data leakage.
✅ AMPC's Solution
Data is encrypted and distributed across multiple independent institutions. Unless all institutions collude, it's impossible to reconstruct the original data.
2. How It Works
Data Flow
Processing at the Orb
Iris data is cryptographically processed directly inside the Orb and anonymized. The original iris image never leaves the Orb.
Splitting into Shards
The anonymized data is split into multiple encrypted "shards" (fragments). Each shard is meaningless on its own.
Distributed Storage
Each shard is sent and stored on computation nodes at different partner institutions with end-to-end encryption.
Collaborative Computation
Duplicate checks (verifying the same person hasn't registered twice) are performed collaboratively by institutions without revealing their shards. This is the core of "Multi-Party Computation."
🧩 Shard Analogy
Imagine splitting a safe's key into 5 parts and giving each part to a trusted person. The safe can only be opened when all 5 come together. AMPC goes further—it's an innovative technology that lets each party "verify the safe's contents" while keeping their parts separate, without ever gathering together.
3. Quantum Computer Resistance
The Quantum Computer Threat
When powerful quantum computers become reality, many current encryption methods could be broken. This creates the risk of "harvest attacks"—collecting data now to decrypt it later.
AMPC's Quantum-Safe Measures
Quantum-Resistant Cryptography
Uses next-generation cryptographic algorithms that are difficult to break even with quantum computers
Upgradeable Design
Flexibility to upgrade the system when stronger encryption is needed in the future
Protection Through Anonymization
Even if shards were decrypted, the data is already anonymized, making individual identification difficult
Resistance Through Distribution
Would require decrypting all partner shards simultaneously, making it practically impossible
4. Partner Institutions
To ensure AMPC's reliability, independent institutions from around the world participate as partners. These institutions are geographically and organizationally independent, minimizing the risk of collusion.
UC Berkeley RDI
University of California, Berkeley's Center for Responsible Decentralized Intelligence. A world-leading institution in blockchain and AI research.
FAU Erlangen-Nürnberg
Friedrich-Alexander University in Germany. One of Europe's leading engineering and computer science research institutions.
Nethermind
A global blockchain and cryptographic security company with extensive experience in the Ethereum ecosystem.
KAIST
Korea Advanced Institute of Science and Technology. A leading science and technology research institution in Asia.
UTEC (Peru)
University of Engineering and Technology, Peru. Participating as a technology research hub in South America.
Why Universities as Partners?
Universities are not bound by commercial interests and maintain academic credibility and independence. By being distributed around the world, they also reduce the risk of interference by any specific government or corporation.
5. Benefits for Users
Minimized Data Breach Risk
Even if one institution is hacked, your data cannot be reconstructed
No Single Point of Failure
Not even World Foundation or TFH can access your data alone
Future-Proof Protection
Your privacy remains protected even when the quantum computing era arrives
Transparency
AMPC is open source, and anyone can verify the code
6. Technical Details
GPU High-Speed Computing
The AMPC protocol is implemented on GPUs using NVIDIA CUDA, enabling approximately 50 million pairwise comparisons per second.
This enables fast duplicate checks even with a user base of billions.
Secret Sharing Scheme
Applies cryptographic techniques such as Shamir's Secret Sharing. Data is split into multiple shares, and cannot be reconstructed without a threshold number of shares.
End-to-End Encryption
Data is encrypted at the Orb before being sent to each partner. It remains encrypted at all times—both in transit and at rest.
Open Source
The AMPC implementation is planned to be released as open source, enabling audits and improvements by security researchers.
📚 Learn More
For technical details, please refer to the official World blog post: "Introducing AMPC: Another leap in privacy and performance for World ID".
Get bonus WLD with this invite code
Use this invite code to receive additional WLD.
Get World App6YQ47NH