Before 2020 ended, IBM Security launched a service that allowed its clients to analyze data on the cloud or third-party platforms without ever decrypting them. With fully homomorphic encryption (FHE), companies are given a testing environment used in developing prototype applications without exposing confidential data.
Current encryption techniques protect data during storage or in transit but creates a window of opportunity that leaves data vulnerable to exposure and theft when it is being decrypted. Although FHE is still an emerging technology, it is already closing the gaping hole present in the majority of encryption solutions today.
During an online briefing and demo, IBM Research AI Security group manager Omri Soceanu shared how FHE changes the way enterprises keep data secure. He explained that the process includes not just being able to encrypt the data in transit, but also perform computations without decrypting the data as well. If the user’s server or cloud account is ever accessed by malicious entities, these threat actors would not be able to see the data itself.
“If you wanted to perform computation over that data, then you would have to decrypt it. If a malicious entity got access to your server or your cloud account, at that exact point when the data is vulnerable, they would be able to see the data and steal it. FHE changes this,” Soceanu said.
Before companies are able to take full advantage of what FHE as to offer, they must first be able to learn and experiment with it, the tech giant said. With the IBM Security Homomorphic Encryption Services, businesses are provided with a scalable hosting environment on the IBM Cloud.
FHE is designed to be resistant against future quantum-computing needs, based on “quantum safe” lattice cryptography. Before being able to perform at seconds per bit, initial FHE computations were too slow to be viable for everyday usage. Now, research and advisory firm Gartner estimates that by 2025, at least 20% of companies will have budget allocated for projects revolving around FHE, unlike today where it is less than 1%.
“We have to accept the fact that every organization will use multiple clouds – either what they have internally in their data center or what they consume from the public cloud space from any of the popular vendors. Sharing data and making data available for compute somewhere else is a pattern that we will have now after Covid,” added Patrick Bruinsma, IBM Z client technical manager.
Moving forward, IBM said it will be working closely with its clients in developing prototype solutions and use cases on FHE technology, initially focusing on developers and crypto engineers. Some of the first use cases include analyzing encrypted data and training AI and machine learning models and maintaining privacy and confidentiality controls at the same time.
