Proposal Explorer

Add claims to see which ones the mechanisms address.

In production

• R3 Model identity attestation for showing users that a service runs the declared model weights
• R3 Sampled inference recomputation for checking untrusted workers' activations against the declared model, prompt and precision

Built for an adversarial prover

Side-channel suppression for isolated facilities and Sampled inference recomputation

No new hardware needed

Model identity attestation and Sampled inference recomputation

Flaws since mitigated

• Launch-state attestation does not by itself cover weights loaded later in Model identity attestation ^{2 14}

Published attempts to break a system, including those that found failures. Testing history does not show that open flaws are resolved.

• Side-channel suppression for isolated facilities: Analysis
• Model identity attestation: Independent red-team
• Sampled inference recomputation: Analysis

Open critical flaws

• Underlying attestation can be forged or relayed in Model identity attestation

  Critical for the enclave route against an operator with physical access to affected hardware, or control of an unpatched SEV-SNP hypervisor. It does not apply to the recomputation route. PAL*M excludes physical attacks, and Tinfoil acknowledges this boundary.

  The enclave route inherits the platform-specific TEE attestation failures. Intel TDX forgery and H100 relay were demonstrated with physical access and host control. Battering RAM defeated AMD SEV-SNP attestation on DDR4 servers; RMPocalypse did so from malicious host software on platforms without AMD's fixes. These demonstrate failures of the trust roots, not of each model-commitment protocol. ^{3 5 9 11 12 13}

  Response: The TEE.fail authors report that physical interposer attacks are outside Intel's and AMD's threat models. AMD reports fixes for RMPocalypse.

  Demonstrated attack · Critical · Open · Inherited finding. On the record · Related finding in TEE remote attestation for AI workloads

  Related mechanism R1 Hardware-enabled guarantees (flexHEG) and guarantee processors: A tamper-protected enclosure around the chip is the proposed answer when the party that holds the hardware may attack it physically. A pointer, not evidence that this flaw is mitigated. Add

Open significant flaws

8 flaws in 3 mechanisms

• Supply-chain implants may evade inspection in Side-channel suppression for isolated facilities

  Cankaya identifies malicious hardware embedded deep in purchased components as a residual risk that visual inspection and disassembly may not catch. He notes that radiographic examination under high-security standards could mitigate it. ¹

  Theoretical argument · Significant · Open. On the record

• Openings for airflow, power and optics weaken shielding in Side-channel suppression for isolated facilities

  Cankaya notes that keeping attenuation high while passing high-power airflow, cabling and optical links adds complexity beyond existing shielded-enclosure specifications. ¹

  Theoretical argument · Significant · Open. On the record

• Inspection assumptions may not hold in Side-channel suppression for isolated facilities

  The design's statistical argument assumes that visual or disassembly inspection catches every flaw that is present in a sampled unit. Cankaya is unsure whether destructive teardowns are defence-dominant or offence-dominant. ¹

  Open question · Significant · Open. On the record

• For private models, a user can confirm consistency but not content in Model identity attestation

  When weights are not published, users can check that the same root hash is served each time, but not what the model is. Pairing the hash with an attested evaluation, as in Attestable Audits, is one proposed remedy. ^{2 15}

  Open question · Significant · Open. On the record

• Recomputation depends on trusted logging and randomness, and its tolerance leaves a covert channel in Model identity attestation

  The recomputation variant assumes that every input, output and seed is logged correctly, and that the attacker can neither predict nor manipulate which messages are sampled for verification. Legitimate nondeterminism concentrates at a few token positions, and slow leaks within the tolerated slack remain possible. An independent study showed that an adversary who controls the prompts roughly doubles the bits leaked per token, reducing the exfiltration slowdown from 146–254 times under benign prompts to 60–118 times. The attack targets the exfiltration bound, not the check that outputs match the declared model. ^{4 10}

  Demonstrated attack · Significant · Open. On the record

  Related mechanism R1 Network taps and certifiers: Taps are proposed to copy and hash traffic on the monitored links, reducing reliance on the prover's own log. This still depends on the monitored boundary and trusted capture. A pointer, not evidence that this flaw is mitigated. Add

  Related mechanism R3 Deterministic and bit-exact inference: Bit-exact inference would remove the numerical tolerance that leaves this channel. A pointer, not evidence that this flaw is mitigated. Add

• Tolerance for numerical noise leaves a covert channel in Sampled inference recomputation

  Schemes that accept approximate matches can put an upper bound on an adversary's covert bandwidth, but they cannot close the channel. The weight-exfiltration detector cut exfiltratable information to under 0.5%, not to zero, on a 30-billion-parameter mixture-of-experts model under benign prompt traffic. Its authors called the channel's size under adversarial prompts an open empirical question. An independent study showed that an adversary who controls the prompts roughly doubles the bits leaked per token. Across six models, that cut the slowdown from 146–254 times under benign prompts to 60–118 times. The attack widens the exfiltration bound. It does not target the check that outputs match the declared model. ^{4 10 25}

  Demonstrated attack · Significant · Open. On the record

  Related mechanism R3 Deterministic and bit-exact inference: Bit-exact inference would remove the numerical tolerance if exact replay can be deployed with the required weights and configuration. A pointer, not evidence that this flaw is mitigated. Add

• Only recorded traffic is checked in Sampled inference recomputation

  Recomputation checks that recorded, declared workloads are correct. It cannot show that the record is complete. The published schemes do not cover hidden workloads run on the same compute, or substituted work. Rinberg et al. say their exfiltration-detection scheme cannot stand alone. ^{4 26}

  Theoretical argument · Significant · Open. On the record

  Related mechanism R1 Network taps and certifiers: Taps copy and hash all traffic on the monitored links, which bears on whether the traffic record is complete. They do not show what else ran on the same chips. A pointer, not evidence that this flaw is mitigated. Add

• Some inference optimizations are not covered in Sampled inference recomputation

  TOPLOC's authors state that it cannot detect speculative decoding in which a cheaper model does the decoding. They did not test whether it distinguishes types of key-value (KV) cache compression. DiFR was evaluated only on sampling from a single model. Its authors sketch an extension to one speculative-decoding algorithm but do not test it. ^{16 17}

  Theoretical argument · Significant · Open. On the record

Open minor flaws

1 mechanism with minor findings

• Mixed hardware widens the honest baseline in Sampled inference recomputation

  When honest reference runs span different GPU types, the spread of benign scores grows. In DiFR's tests on Qwen3-30B-A3B, pooling A100 and H200 runs left Token-DiFR unable to separate the two smallest tested changes, a temperature of 1.1 instead of 1.0 and a simulated top-2 sampling bug, at the target false-positive rate, while cross-entropy separated them. Matched provider and verifier environments, or pooling that weights rare large deviations, restored detection. ¹⁶

  Open question · Minor · Open. On the record

Not yet demonstrated

R1 Side-channel suppression for isolated facilities

Mechanisms on the map that are not in the proposal. Pointers, not recommendations. ?

• R1 Hardware-enabled guarantees (flexHEG) and guarantee processors Add
  • Bears on the open critical flaw “Underlying attestation can be forged or relayed” in Model identity attestation. A tamper-protected enclosure around the chip is the proposed answer when the party that holds the hardware may attack it physically.
• R3 Deterministic and bit-exact inference Add
  • Bears on the open significant flaw “Recomputation depends on trusted logging and randomness, and its tolerance leaves a covert channel” in Model identity attestation. Bit-exact inference would remove the numerical tolerance that leaves this channel.
  • Bears on the open significant flaw “Tolerance for numerical noise leaves a covert channel” in Sampled inference recomputation. Bit-exact inference would remove the numerical tolerance if exact replay can be deployed with the required weights and configuration.
  • Model identity attestation waits on it. Numerical nondeterminism limits how tightly recomputation can pin down the model and sampling.
• R1 Network taps and certifiers Add
  • Bears on the open significant flaw “Recomputation depends on trusted logging and randomness, and its tolerance leaves a covert channel” in Model identity attestation. Taps are proposed to copy and hash traffic on the monitored links, reducing reliance on the prover's own log. This still depends on the monitored boundary and trusted capture.
  • Bears on the open significant flaw “Only recorded traffic is checked” in Sampled inference recomputation. Taps copy and hash all traffic on the monitored links, which bears on whether the traffic record is complete. They do not show what else ran on the same chips.
  • Sampled inference recomputation waits on it. In tap-based retrofit designs, recording all inference traffic needs network taps and recomputation servers that can ingest it, in the worst case one recomputation-server network interface per inference front-end interface.
• R3 TEE remote attestation for AI workloads Add
  • Model identity attestation waits on it. Attestation that resists physical attackers, for the enclave variant.

Blockers

10 blockers recorded

• Side-channel suppression for isolated facilities
  • No prototype or red-team exists; the design is a first-pass viability study. Adversarial validation ¹
  • Volume costs of TEMPEST-grade power-line filters are uncertain, because existing products are mostly made to order. Performance & compatibility ¹
• Model identity attestation
  • Attestation that resists physical attackers, for the enclave variant. Hardware trust. Waits on TEE remote attestation for AI workloads ⁹
  • Numerical nondeterminism limits how tightly recomputation can pin down the model and sampling. Protocol soundness. Waits on Deterministic and bit-exact inference ⁴
  • The recomputation variant needs the verifier to hold the declared weights. Access & governance ⁴
• Sampled inference recomputation
  • In tap-based retrofit designs, recording all inference traffic needs network taps and recomputation servers that can ingest it, in the worst case one recomputation-server network interface per inference front-end interface. Coverage & hidden compute. Waits on Network taps and certifiers ^{22 28}
  • In retrofit designs, the recomputation server must sit inside the prover's data centre, possibly under the prover's physical control, and still be protected from a compromised provider, which Amodo rates 'not on track'. Hardware trust ^{4 22}
  • No independent red-team of a recomputation consistency check has been published (the one independent attack study targets the weight-exfiltration bound), and Amodo rates recomputation red-teaming 'not started'. Adversarial validation ^{10 22}
  • Tolerance-based checks need calibration on trusted hardware and exact knowledge of the provider's sampling procedure, and in one prototype a sampling-implementation mismatch produced large spurious differences. Performance & compatibility ^{16 21}
  • The verifier needs the model weights, so checking a closed-weights model requires a trusted, confidential recomputation environment, which the retrofit designs place inside the prover's facility. Privacy & leakage ^{16 26 27}

From the family or selected implementation's record. ?

Model weights

Depends on the design for Model identity attestation and Sampled inference recomputation.

Not involved: Side-channel suppression for isolated facilities.

Inputs and outputs

Depends on the design for Model identity attestation and Sampled inference recomputation.

Not involved: Side-channel suppression for isolated facilities.

Training data

Not involved: Side-channel suppression for isolated facilities, Model identity attestation and Sampled inference recomputation.

Side-channel suppression for isolated facilities

• R1 AI 2040 inference-only verification stack Proposed architecture, AI Futures Project
• R1 Low-trust AI compute verification system overview Proposed architecture, Machine Intelligence Research Institute
• R1 RAND secure inference data center (SIDC) design Proposed architecture, RAND

Model identity attestation

• R2 Attestable Audits Research prototype, University of Cambridge
• R2 PAL*M Research prototype, University of Waterloo
• R3 Tinfoil model identity (Modelwrap) Product, Tinfoil

Sampled inference recomputation

• R1 AI 2040 inference-only verification stack Proposed architecture, AI Futures Project
• R2 DiFR (Divergence From Reference) Research prototype
• R1 Low-trust AI compute verification system overview Proposed architecture, Machine Intelligence Research Institute
• R1 SASH confidential network logger Research prototype, Singapore AI Safety Hub (SASH)
• R3 TOPLOC Open-source project, Prime Intellect

1. Suppressing Side Channels in an Untrusted Data Center via Retrofitted Defenses, N. Cankaya (2026). Original
2. How Tinfoil Proves Exactly What Model Is Running, Tinfoil Team (2026). Original
3. PAL*M: Property Attestation for Large Generative Models, P. Chantasantitam et al. (2026). Original
4. Verifying LLM Inference to Detect Model Weight Exfiltration, R. Rinberg et al. (2025). Original
5. A primer on secure enclaves, Tinfoil (2026). Original
6. Backend infrastructure, Tinfoil (2026). Original
7. How verification works in Tinfoil, Tinfoil (2026). Original
8. modelwrap: Reproducible dm-verity read-only image of Huggingface models, Tinfoil (2026). Original
9. TEE.fail: Breaking Trusted Execution Environments via DDR5 Memory Bus Interposition, J. Chuang et al. (2026). Original
10. Adversarial Entropy Inflation Against Gumbel-Based Inference Verification, N. Kezins (2026). Original
11. Battering RAM: Low-Cost Interposer Attacks on Confidential Computing via Dynamic Memory Aliasing, J. De Meulemeester et al. (2026). Original
12. RMPocalypse: How a Catch-22 Breaks AMD SEV-SNP, B. Schlüter & S. Shinde (2025). Original
13. SEV-SNP RMP Initialization Vulnerability (AMD-SB-3020), AMD (2025). Original
14. On TEEs for Privacy-Preserving Monitoring in AI Governance, Gloria Z (2026). Original
15. Attestable Audits: Verifiable AI Safety Benchmarks Using Trusted Execution Environments, C. Schnabl et al. (2025). Original
16. DiFR: Inference Verification Despite Nondeterminism, A. Karvonen et al. (2025). Original
17. TOPLOC: A Locality Sensitive Hashing Scheme for Trustless Verifiable Inference, J. M. Ong et al. (2025). Original
18. PrimeIntellect-ai/toploc (GitHub repository), Prime Intellect (2025). Original
19. INTELLECT-2: A Reasoning Model Trained Through Globally Decentralized Reinforcement Learning, Prime Intellect Team et al. (2025). Original
20. adamkarvonen/difr (GitHub repository), A. Karvonen (2025). Original
21. Scaling Recomputation Inference Verification, Amodo Design (2026). Original
22. AI 2040 Plan A — Verification SITREP, Amodo Design (2026). Original
23. SYNTHETIC-2 Release: Four Million Collaboratively Generated Reasoning Traces, Prime Intellect (2025). Original
24. An Inference Verification Prototype — Stage 1, Amodo Design (2026). Original
25. Bit-Exact AI Inference Verification Without Performance Tradeoffs, N. Cankaya (2026). Original
26. Example Schemes for Verifying High-Stakes AI Agreements, Amodo Design (2026). Original
27. A System Overview for Near-Term, Low-Trust AI Compute Verification, N. Cankaya (2026). Original
28. Verification Plan, R. Dean (2026). Original