Advancing Responsible AI — Assessing Fairness in AI Models

Our groundbreaking system, in collaboration with Meta AI, enables privacy-preserving assessment of AI models’ fairness and bias, fostering responsible AI practices while upholding data privacy.

The designed system uses a combination of MPC, Homomorphic Encryption, Differential Privacy, and Zero-Knowledge Proofs, to achieve the best privacy protection while still being able to do the computation at a scale of millions of users.