Open for AI: Secure paths to data collaboration volume and diversity
summary
Overview of concrete capabilities and results: hardware TEEs like Intel SGX and virtualization TEEs like TDX run in Azure (SGX GA in 17 regions, TDX in limited preview) and support attested keys and Project Amber attestation, Gramine lets unmodified containers use SGX, and OpenFL drives federated learning across 71 global sites yielding a reported 27% accuracy lift; also mentioned are Kata Confidential Containers, Intel Getty demoed for vision models, and model/IP protection via attestation and CCF governance.
AI-written summary · about
chapters (20)
- 0:00Intro
- 1:28What are developers missing
- 5:55Microsoft and Intel
- 8:45Intel and Software
- 13:58Federated Learning
- 16:25Use Cases
- 24:54The Beekeeper AI
- 26:38Intel Trust Domain Extensions
- 29:13attestation
- 31:19containers
- 33:14Future of Security
- 34:36Companies will continue to innovate
- 35:32Data confidentiality
- 37:30Public transit AI recommendations
- 38:59What are we trying to solve
- 40:00Can using Openfl help keep my model secure
- 41:07Future of CCF
- 42:16Join CCF
- 43:16GDPR
- 44:08Wrap up
related talks
37mConfidential AI: Enabling Trustworthy Ecosystems in the Age of Agentic Intelligence - Session
05mSponsored Keynote: Bringing Confidential Computing to Kubernetes
36mAttacking & Defending Kubernetes TEE Enclaves in Critical Infrastructure
31mProject Amber - Intel's attestation service
20mKeynote: Future-Proofing Confidential AI
24mStrictly Confidential: Use Cases & Demos of AI Workflows in TEEs
28mGenerative AI + Proprietary Data
42mSponsored Session: Confidential Computing - New Capabilities for New Workloads