01 / The mechanism and its boundary
What is being described
Attestable says it develops zero-knowledge proofs for AI systems S-3602. It has published the following on verification:
- Attestable reports a zero-knowledge prover for production-scale transformer inference whose security rests only on cryptographic hash functions S-1101. See Attestable zero-knowledge inference prover and zero-knowledge proofs of inference.
- It proposes combining inference proofs with proof-of-work accounting, and notes that the verifier "needs a credible estimate of the compute available" to the actor S-1102; see Proofs of useful work for capacity accounting.
- It proposes a verification firewall at the external interfaces of an environment that holds model weights, which would select outputs for proof at random after each output is committed S-1103.
Connections in the research map
Related research
Sources and provenance
- S-3602 / Tier B
Attestable homepage ↗
· 2026 · Attestable
Supports: development of zero-knowledge proofs for AI systems
Version and catalogue details - S-1101 / Tier C
Proving LLMs at Scale ↗
Attestable · 2026 · Attestable blog
Supports: zero-knowledge prover for production-scale transformer inference; hash-only security (provider-reported)
Version and catalogue details - S-1102 / Tier C
Pacing AI Requires Proof ↗
Attestable · 2026 · Attestable blog
Supports: proposal to combine inference proofs with proof-of-work accounting; need for a compute estimate
Version and catalogue details - S-1103 / Tier C
From Verifiability to Model-Weight Security ↗
Attestable · 2026 · Attestable blog
Supports: verification-firewall proposal with randomly selected proofs
Version and catalogue details
- Source review date
- 2026-09-25
- Drafted by (source map)
- ai
- Review handles (source map)
- codex-review