Secure Computation

TECHNICAL DEFINITION

A cryptographic primitive, often called Secure Multi-Party Computation (SMC or MPC), that allows multiple parties to jointly compute a function over their private inputs without revealing any individual input to the other parties, ensuring data confidentiality.

BACKGROUND

Generative artificial intelligence (GenAI) is a subfield of artificial intelligence (AI) that uses generative models to generate text, images, videos, audio, software code or other forms of data. These models learn the underlying patterns and structures of their training data, and use them to generate new data in response to input, which often takes the form of natural language prompts.

SYNONYMS & ALIASES

• Multi-party computation (MPC)
• SMC
• privacy-preserving computation

USAGE NOTE

Secure computation can be used in scenarios like joint financial analysis where parties want to keep their individual data private.

DEVELOPERS

Organizations developing technology related to Secure Computation.

• Intel

  Intel develops hardware-based security technologies like Intel SGX (Software Guard Extensions) and contributes to confidential computing, enabling secure execution environments for AI models and data processing.

• Microsoft

  Microsoft offers Azure Confidential Computing services and invests heavily in research and development of privacy-preserving technologies, including homomorphic encryption and multi-party computation for secure AI workloads.

• Google

  Google Cloud provides confidential computing instances and conducts extensive research in privacy-preserving AI, focusing on technologies like federated learning, secure aggregation, and differential privacy, which leverage secure computation principles.

• IBM

  IBM offers Hyper Protect services that leverage Trusted Execution Environments (TEEs) and conducts deep research in cryptography, including secure multi-party computation (MPC) and homomorphic encryption, applicable to secure AI and data processing.

• Zama

  Zama specializes in homomorphic encryption (HE), developing open-source libraries and solutions that allow computations on encrypted data, which is crucial for building privacy-preserving AI systems without decrypting sensitive information.

• Inpher

  Inpher provides secure multi-party computation (MPC) platforms that enable organizations to collaborate and perform analytics and machine learning on sensitive data without exposing the underlying raw data to any party.

• Decentriq

  Decentriq offers a confidential computing platform that uses Trusted Execution Environments (TEEs) and multi-party computation (MPC) to enable secure data collaboration and AI model training, ensuring data privacy and compliance.