Practical Accountability of Secret Processes
summary
Federal electronic-surveillance orders get cryptographic commitments posted to a tamper‑resistant ledger, companies and agencies post zero‑knowledge SNARK proofs tying requests to orders, and courts compute aggregates via hierarchical MPC that splits each court’s data into 12 secret shares. Prototype uses JIFF for MPC and libsnark for proofs; flat MPC needs ~7 minutes for 300 judges while hierarchical sums ~1,000 judges in ~20 seconds; existential SNARKs are ~287 bytes, proving keys hundreds of MB, proving seconds–2 minutes, verify in <1s. Collusion remains an unresolved limitation.
AI-written summary · about
chapters (9)
- 0:00The problem of secrecy
- 1:31Surveillance process
- 3:22Accountability challenges
- 5:04Cryptographic proposal
- 6:43System design
- 9:54Hierarchical MPC
- 12:29Practical performance
- 15:01SNARKs evaluation
- 18:57Conclusion and Q&A
related talks
32mPublic Accountability vs Secret Law: Can They Coexist
18mzkSaaS: Zero-Knowledge SNARKs as a Service
01h 08mSecure Computation 1
34mSNARKs and their Practical Applications
40mCryptography in Distributed Systems
56mCollaborative zk-SNARKs: Proving as One over Distributed Secrets
51mDistributed Zero-Knowledge Proofs and Applications to Secure Computation
25mHigh-Speed zkSNARKs without Trusted Setup