Fortanix Introduces Confidential AI

Imagine your Artificial Intelligence (AI) churning in public cloud services -- processing lots of sensitive data such as Personally Identifiable Information (PII) or even genomics data. This is the challenge being addressed by a new crop of security startups, including Fortanix.

A new approach to cloud security, referred to as confidential computing or confidential cloud, uses several layers of security, including identity-based security as well as hardware-level and memory-level encryption for cloud services.

The goal is to lock down not just "data at rest" or "data in motion," but also "data in use" -- the data that is being processed in a cloud application on a chip or in memory. This requires additional security at the hardware and memory level of the cloud, to ensure that your data and applications are running in a secure environment.

What Is Confidential AI in the Cloud?

One of the companies working in this area is Fortanix, which has announced Confidential AI, a software and infrastructure subscription service designed to help improve the quality and accuracy of data models, as well as to keep data models secure.

According to Fortanix, as AI becomes more prevalent, end users and customers will have increased qualms about highly sensitive private data being used for AI modeling. Recent research from Gartner says that security is the primary barrier to AI adoption.

One customer using the technology pointed to its use in locking down sensitive genomic data for medical use.

“Fortanix is helping accelerate AI deployments in real world settings with its confidential computing technology,” said Glen Otero, Vice President of Scientific Computing at Translational Genomics Research Institute (TGen). "The validation and security of AI algorithms using patient medical and genomic data has long been a major concern in the healthcare arena, but it's one that can be overcome thanks to the application of this next-generation technology."

Creating Secure Hardware Enclaves

One of the goals behind confidential computing is to develop hardware-level security to create trusted and encrypted environments, or enclaves. Fortanix utilizes Intel SGX secure enclaves on Microsoft Azure confidential computing infrastructure to provide trusted execution environments.

Secure enclaves are one of the key elements of the confidential computing approach. Confidential computing protects data and applications by running them in secure enclaves that isolate the data and code to prevent unauthorized access, even when the compute infrastructure is compromised.

In cloud applications, security experts believe that attack patterns are increasing to include hypervisor and container-based attacks, targeting data in use, according to research from the Confidential Computing Consortium. In this case, protecting or encrypting data at rest is not enough. The confidential computing approach strives to encrypt and limit access to data that is in use in an application or in memory.

Fortanix, along with several other startups including Anjuna, Anqlave, and CYSEC – just to name a few – are members of the Confidential Computing Consortium, a group of companies targeting confidential computing. More than 20 industry leaders have joined the group, including Alibaba, Anjuna, ARM, Baidu, CYSEC, Facebook, Fortanix, Google Cloud, IBM, Intel, Microsoft, Oracle, Red Hat, Tencent, and VMware.

Fortanix was also featured in this year's Futuriom 40 report of leading cloud technology startups.