Confidential AI at Intel

Generative artificial intelligence (AI) tools and large language models (LLMs) have exploded onto the marketplace, allowing businesses to optimize workflows, streamline processes and become more efficient on a global scale.

However, as companies embrace this technology, they become more aware of how much this data processing impacts their Zero Trust policies to prevent exposing sensitive, proprietary or confidential data and their compliance requirements following recently introduced regulations such as the European Union’s AI Act and the U.S. Executive Order on the Safe, Secure and Trustworthy AI. The inherent value of AI models themselves also warrants protection; intellectual property – such as bespoke algorithms and LLMs – are the result of years of research and development and millions of dollars of investment.

Confidential AI helps protect this data and can enable companies to continue harnessing the power of AI while meeting the security, privacy and compliance standards required to do business. It also protects proprietary generative models from exposure, securing valuable intellectual property.

What is Confidential AI?

Confidential AI exists at the intersection of artificial intelligence (AI) and confidential computing, bridging the gap between Zero Trust policies designed to secure private data and generative AI, which often relies on cloud compute power to be trained and process complex tasks and requests. For businesses to trust in AI tools, technology must exist to protect from exposure inputs, trained data, generative models and proprietary algorithms. Confidential AI helps make that happen.

Confidential AI utilizes confidential computing principles and technologies to help protect data used to train LLMs, the output generated by these models and the proprietary models themselves while in use. Through vigorous isolation, encryption and attestation, confidential AI prevents malicious actors from accessing and exposing data, both inside and outside the chain of execution.

Intel’s Approach to Confidential AI

AI will only be genuinely accessible to everyone when it is built ethically and responsibly. Intel collaborates with technology leaders across the industry to deliver innovative ecosystem tools and solutions that will make using AI more secure while helping businesses address critical privacy and regulatory concerns at scale.

Confidential AI – Protecting Data and Models with Intel Confidential Computing

Intel builds platforms and technologies that drive the convergence of AI and confidential computing, enabling customers to secure diverse AI workloads across the entire ecosystem. Intel offers the most comprehensive confidential computing portfolio in the industry today:

• Application Isolation with Intel® Software Guard Extensions (Intel® SGX)
• Virtual Machine Isolation with Intel® Trust Domain Extensions (Intel® TDX)
• Independent Trust Attestation Services through Intel® Trust Authority

 Real World Impact

Intel’s innovative and comprehensive approach to confidential computing and confidential AI offers benefits to industries that rely on storing and processing sensitive information, such as healthcare, government, finance, retail and others. With confidential AI, businesses can rapidly process large volumes of data through their training models while maintaining higher levels of security and compliance.

Read on for how Intel customers utilize Intel technologies and confidential AI:

• Case Study: Ant Group Selects Alibaba Cloud with Intel for Confidential AI
• Case Study: Equideum Health: Revolutionizing Health Data
• InTechnology Podcast: Azure CTO Talks Confidential Computing and Confidential AI
• Video: Federated Learning, a New Model for Confidential Computing
• Case Study: Accelerated AI Inference with Confidential Computing
• Case Study: Securely Use Confidential Data with Intel® Software Guard Extensions and Fortanix Confidential AI
• Case Study: Revolutionizing Healthcare with BeeKeeperAI
• Whitepaper: Privacy-Preserving Data Collaboration Methods that Accelerate Healthcare Innovation

• Interview on Confidential AI with Intel’s Anand Pashupathy (AIThority)
• Confidential AI: Enabling Secure Processing of Sensitive Data (Help Net Security)
• Confidential AI: Intel Seeks to Overcome AI’s Data Protection Problem (Think Digital Partners)
• The Importance of Protecting AI Models (Inside Big Data)
• Intel CTO Greg Lavender on Confidential AI: The Convergence of Security, Privacy & AI
• Security as a Business Enabler: Unlocking Opportunities with Confidential Computing
• Confidential AI Protects Data and Models Across Clouds (Dark Reading)
• Intel CEO Pat Gelsinger’s ‘AI Everywhere’ Keynote
• InTechnology Podcast: AI and Cybersecurity for Open Source
• Seamless Attestation of Intel® TDX and NVIDIA H100 TEEs with Intel® Trust Authority
• Intel, Nvidia Collaborate to Deliver Confidential AI Solutions that Strengthen AI Security, Privacy
• InTechnology Podcast: Security Trends in AI and Confidential Computing
• Intel CTO Greg Lavender on Confidential Computing: From Niche to Mainstream
• Intel: Protecting Data and Models within Emerging AI Workflows (Moor Insights & Strategy Analyst Whitepaper)
• BeeKeeperAI Secures AI Algorithms with Infrastructure from Intel, Microsoft and Fortanix (IDC Analyst Whitepaper)
• Intel’s Transition of OpenFL Primes Growth of Confidential AI