Fully Homomorphic Encryption
Privacy enhancing technologies (PETs) are rapidly changing the way data is shared and protected across sectors such as healthcare, law enforcement and finance. By enabling information sharing without revealing personal details, PETs can facilitate collaboration and create new opportunities for data-driven insights. PETs are based on groundbreaking concepts such as secure multi-party computation, zero-knowledge proofs, and fully homomorphic encryption. Originally considered academic exercises, these theories have evolved to address real-world privacy challenges and are now being adopted by large technology companies.
I have been looking at Fully Homomorphic Encryption (FHE) - a (breakthrough) encryption technology that allows computations to be performed on encrypted data without ever decrypting it. #FHE has the potential to change the way we think about privacy and security, especially in industries where data is sensitive and confidential and where our customers want to be sure that their data is kept private and secure.
I played with the concrete.python library (disclaimer - I know the founder) and was able to apply encryption - analysis - decryption with relative ease. It's still (relatively) slow on my laptop, based on a simple benchmarking proof of concept (5 minutes for 7-bit operations, 1h+ for 8-bit operations), but it shows we can do it. Their Kaggle example is quite interesting for those who want to try it ( https://www.kaggle.com/code/concretemlteam/titanic-with-priv... ).
Will this be the change of tomorrow? Certainly not, but just as certainly PETs will change the way we work with data!