NEWS · COMPANIES · #877
Google adds secure server-side persistent memory to Private AI Compute
Google (with DeepMind and multiple internal teams) announced a technical update to its Private AI Compute platform that enables persistent, cross-device server-side memory using hardware-enforced secure enclaves and device-held encryption keys; the company published an updated technical brief, a tamper-proof public record of server software, and results of an independent security audit for verification. The design keeps cryptographic keys on users' devices while data is stored encrypted in isolated cloud enclaves and is briefly decrypted only inside the enclave to service requests.
KEY POINTS
- Google (with DeepMind and multiple internal teams) announced a technical update to its Private AI Compute platform that enables persistent, cross-device server-side memory using hardware-enforced secure enclaves and device-held encryption keys; the company published an updated technical brief, a tamper-proof public record of server software, and results of an independent security audit for verification.
- The design keeps cryptographic keys on users' devices while data is stored encrypted in isolated cloud enclaves and is briefly decrypted only inside the enclave to service requests.
- This enables cloud-scale, continuous personal AI assistance while aiming to preserve on-device-grade privacy by keeping keys off the server and providing public verification of server software and an independent audit.
WHY IT MATTERS
This enables cloud-scale, continuous personal AI assistance while aiming to preserve on-device-grade privacy by keeping keys off the server and providing public verification of server software and an independent audit.