Presentation Material

AI Generated Summary^{may contain errors}

The speaker is discussing the challenges of implementing homomorphic encryption, which allows computations to be performed on encrypted data without decrypting it first. The main concerns are:

1. Data size and storage: Encrypted data is very large, making storage difficult, especially when dealing with millions of records.
2. Speed: As the size of the data increases, processing speed decreases significantly, making it a major constraint.

The speaker cites an example from IBM, where their homomorphic encryption model took 11.5 days to process encrypted data, compared to 1 second for clear text.

Additionally, there are two more issues:

3. Integrity: Homomorphic encryption does not provide integrity, as people can still tweak the algorithm and get undesired results.
4. Lack of formalized standards: There is no standardized framework for building homomorphic encryption algorithms, making it difficult to ensure security and regulatory compliance.

To address these challenges, researchers are working on:

1. Integrating homomorphic encryption with confidential computing: This aims to manage the integrity factor by enclaving the entire processing into a secure environment.
2. Reducing data size and increasing speed: Multiple patents have been filed to improve the efficiency of homomorphic encryption.

The speaker concludes that while homomorphic encryption is not yet commercially adoptable due to these challenges, there is significant activity happening in this space, and it may take 5-10 years for it to reach a state where it can be widely adopted.