Fully Homomorphic Encryption (FHE) and Applications

FHE is like having a secret party where everyone wears masks, and even the DJ can't see who's dancing!

Introduction

Fully Homomorphic Encryption (FHE) is technology that enables compute over data without decrypting it. It enables Data processing, AI and ML workflows over data in the cloud and blockchains without ever decrypting it, hence, users’ data is always protected.

To better understand, if google drive allows you to search across your documents and image stored, it would do it while decrypting your data on cloud and running algorithms on it. So, while your data is encrypted in transit while uploading and fetching data from google drive but its fully public on google cloud’s machines.

FHE solves this by enabling an encryption scheme where your data is not just uploaded privately and protected in transit but also in rest and during data processing over it. This solves an important issue of privacy but also for user experience as well. As a user, one might love the ease of use of searching across your documents, but FHE enables now a way to do that without losing privacy of your data. Who wants their private pictures to be part of an upcoming GPT model!

Potential Applications

FHE have major use cases changing the way systems are operating today. Let’s discuss few below -

1. Facial Recognition?- During my gym visit, I have to scan my face when entering the gym, while it helps in easy access and prevents outsiders without membership joining in. I have been concerned on where my face data is stored and what happens if some entity takes over that control and might impersonate people leading to online identity theft. FHE can be used here to securely authenticate self while entering buildings, unlock phones and much more without compromising identity if the provider chooses to incorporate it in their device.
2. Health apps?- The global health app market was valued at approximately $38.2 billion in 2022 and is expected to reach $111.1 billion by 2025, growing at a CAGR of 19.6%. There were over 350,000 health-related apps available in major app stores by the end of 2022. Health apps collect vast amount of users’ data making them attractive target of cyberattacks. FHE can play a big role here protecting user health data and still getting value of these increasingly AI powered health apps for better habits and life improvement.
3. Confidential Smart Contracts?- Use of mixers has proven to be double edge sword for web3 developers keeping transactions history and identity private but inviting strong action by legal authorities for misuse for money laundering, tax evasion and putting developers in jail and sanctioning users. FHE seems to be the only hope for having a compliant way of ensuring privacy on public blockchains like Ethereum, Solana, etc. This is also going to enable use cases around RWAs, Private NFTs and much more.

FHE, MPC and ZK

Can FHE do everything and solve everything alone replacing previous breakthroughs in cryptography like MPC and ZK. Probably No! We will need a combination of these for application specific use cases. Example of how these can come together in an example of running an AI algorithm over a NFT gated Dataset stored on decentralized network like Filecoin -

1. FHE?- Can ensure compute over encrypted data stored on Filecoin without decrypting it
2. MPC?- Encryption has to be done securely and keys have to be stored somewhere. MPC comes here and can be used for threshold decryption
3. ZK?- Zero knowledge proofs can ensure integrity and solve the question of how we trust if intended compute happened over the correct encrypted data

Closing Thoughts

While there are still some challenges with speed and computational intensity of current FHE implementations. Next 6-12 months will be great to see transition towards widely adopted consumer and enterprise products with end-to-end secure applications powered by FHE.

Join Nandit on?X,?LinkedIn?or Email:?[email protected]