Source · Tier B · Technical reportSource · A System Overview for Near-Term, Low-Trust AI Compute Verification
A System Overview for Near-Term, Low-Trust AI Compute Verification
N. Cankaya. 2026. Machine Intelligence Research Institute.
| Original | https://intelligence.org/wp-content/uploads/2026/06/A-system-overview-for-near-term-low-trust-AI-compute-verification.pdf |
|---|---|
| Accessed | 2026-09-25 |
| Authored or published by | Machine Intelligence Research Institute |
| Imported from | hodgkins-ai-verification-papers@c71e59ff0e8e |
| Note | Listed under "Inference verification" in the Hodgkins bibliography (CC BY 4.0). |
Cited by
- R2Bandwidth limits and compartmentalization
- R2Bounding unexplained information in outputs
- R3Deterministic and bit-exact inference
- R1Memory wiping and proofs of secure erasure
- R1Network taps and certifiers
- R1Proofs of useful work for capacity accounting
- R2Safeguard attestation
- R3Sampled inference recomputation
- R1Side-channel suppression for isolated facilities
- R2Tamper evidence for verifier devices
- R2Timed challenge-response and memory-occupation challenges
- R2Zero-knowledge proofs of inference
- R1Low-trust AI compute verification system overview
- R2zkLLM
- Communication between compute groups is bounded
- The declared model is the one being served
- This compute runs inference, not training
- Declared safeguards were applied during inference
- A training run stayed within declared limits
- Model weights have not left the facility
- Compartmentalization
- Cryptographic commitment
- Evidence binding
- FLOP accounting
- Inference and training workloads
- Interconnect bandwidth
- Network tap
- Numerical nondeterminism
- Positive and negative claims
- Proof of space
- Recomputation
- Sampling and assurance
- Side channel
- Threat model
- Undeclared compute
- Verifier
- Weight exfiltration
- Machine Intelligence Research Institute