MemPrivacy Shows a Privacy Layer for AI Memory

越懂你，越危险？MemPrivacy揭示AI记忆下一站

MemTensor and HONOR open-sourced MemPrivacy for edge-cloud agent memory protection using local reversible pseudonymization; MemPrivacy-4B-RL reached 85.97% composite F1 on MemPrivacy-Bench, 50.47 percentage points above OpenAI privacy-filter, while the benchmark covers 200 users and more than 155,000 privacy items.

Why it matters: HKR-H/K/R all pass: the story has a sharp memory-privacy hook, a concrete reversible pseudonymization mechanism, and benchmark numbers. Single-source release from non-frontier labs keeps it at 78.