# AI verification proposal

A proposal built with the Proposal Explorer of the AI Verification Tech Map (https://trustbutveri.fyi/), from its records of 2026-10-09. Interactive version: https://trustbutveri.fyi/explorer/?mechanisms=M-0001,M-0002&implementations=M-0002:I-0012

How to read it: a claim is something one party wants to verify about another's AI hardware or software. A mechanism is a general technique for verifying claims; it is "aimed at" a claim when that is its direct purpose, and "supporting" when it contributes without being aimed at it. A claim is addressed when a mechanism in the proposal is aimed at it and is not excluded by the filters; addressed does not mean verified, so check that mechanism's development status, security evidence and findings. Definitions: https://trustbutveri.fyi/about/methodology/ (roles, properties and findings) and https://trustbutveri.fyi/about/readiness/ (development status).

## Filters

Filters apply to mechanisms only and describe the setting the proposal is for.

None set. Every mechanism on the map was available.

## Overview

One row per mechanism, read from its record. Open failures: critical / significant / minor. The last three columns are the editors' reading of what the verifier sees. Findings are grouped as known failures, scope limitations and open questions. Only known failures count as failures. Counts are an inventory of published findings, not a risk score.

| Mechanism | Development | Security evidence | Prover | Attack testing | Hardware | Open failures | Weights | Inputs and outputs | Training data |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Sampled inference recomputation | Operational use | Published security analysis | Adversarial | Analysis | None | 0 / 2 / 1 | depends | depends | not involved |
| Deterministic and bit-exact inference / Low-trust AI compute verification system overview | Proposed | Published security analysis | Adversarial | Analysis | Retrofit device | 0 / 0 / 1 | unspecified | unspecified | unspecified |

## Claims

No claims chosen.

## Mechanisms

### Sampled inference recomputation

A verifier re-runs a random sample of an AI provider's logged queries on a trusted copy of the declared model and checks the outputs match. ([Sampled inference recomputation](https://trustbutveri.fyi/mechanisms/sampled-inference-recomputation/))

- Assessment: mechanism family.
- Development: Operational use (legacy code R3), assessed for checking untrusted workers' activations against the declared model, prompt and precision.
- Security evidence: Published security analysis. Independent evaluation: unassessed. Formal proof: unassessed. Deployment assurance: unassessed.
- Claims in this proposal: none of them.
- Threat model: adversarial prover. Hardware: none. Prover cooperation: required. Attack testing: analysis. Category: Cryptographic & computational.
- What the verifier sees: model weights depends; inputs and outputs depends; training data not involved. Recomputation needs the weights and sampled requests inside the checking environment. For closed models, the record describes a trusted, confidential environment; disclosure to the verifier depends on that boundary.

### Deterministic and bit-exact inference

A reference design that existing data centres could add: network taps hash all traffic, and air-gapped, independently sourced checkers later re-run randomly chosen records. ([Deterministic and bit-exact inference](https://trustbutveri.fyi/mechanisms/deterministic-inference/))

- Assessment: selected implementation [Low-trust AI compute verification system overview](https://trustbutveri.fyi/implementations/low-trust-compute-verification-system-overview/).
- Development: Proposed (legacy code R1), assessed for screening challenged records to show declared inference compute is not training.
- Security evidence: Published security analysis. Independent evaluation: unassessed. Formal proof: unassessed. Deployment assurance: unassessed.
- Claims in this proposal: none of them.
- Threat model: adversarial prover. Hardware: retrofit device. Prover cooperation: required. Attack testing: analysis. Category: Isolation & system architectures.
- What the verifier sees: model weights unspecified; inputs and outputs unspecified; training data unspecified. This Explorer has no asset-specific exposure assessment for this implementation. Check its source and deployment assumptions.


## Properties

**Operational use**

- Sampled inference recomputation: Operational use (legacy code R3), assessed for checking untrusted workers' activations against the declared model, prompt and precision

**Built for an adversarial prover**

- Sampled inference recomputation
- Deterministic and bit-exact inference

**No new hardware needed**

- Sampled inference recomputation


## Attack testing

Attack testing records published testing for this use. It does not by itself show independent review, a formal proof or that a deployed system is secure.

**Testing history**

- Sampled inference recomputation: Analysis
- Deterministic and bit-exact inference / Low-trust AI compute verification system overview: Analysis


## Limits

**Open significant failures**

- Tolerance for numerical noise leaves a covert channel (known failure, demonstrated attack, in Sampled inference recomputation; https://trustbutveri.fyi/mechanisms/sampled-inference-recomputation/evidence/flaws/1/) [1][9][12]. 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.

  Related mechanism: Deterministic and bit-exact inference (R1, in the proposal). Bit-exact inference would remove the numerical tolerance if exact replay can be deployed with the required weights and configuration.
- Some inference optimizations are not covered (known failure, theoretical argument, in Sampled inference recomputation; https://trustbutveri.fyi/mechanisms/sampled-inference-recomputation/evidence/flaws/3/) [2][3]. 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.

**Family finding context**

- Context for Low-trust AI compute verification system overview. Findings from the mechanism family appear here as context. They apply to an implementation only when its own record lists them, under the conditions stated there. Some kernels remain genuinely nondeterministic (scope limitation, open question, in Deterministic and bit-exact inference; https://trustbutveri.fyi/mechanisms/deterministic-inference/evidence/flaws/1/) [12]. The bit-exact work separates kernels that are deterministic but not batch-invariant from truly nondeterministic ones that use atomic functions. Some integer de-quantization kernels use atomic additions and remain nondeterministic, so exact replay needs backends that avoid them.
- Context for Low-trust AI compute verification system overview. Findings from the mechanism family appear here as context. They apply to an implementation only when its own record lists them, under the conditions stated there. Cross-hardware replay relies on reverse-engineered, closed behaviour (scope limitation, open question, in Deterministic and bit-exact inference; https://trustbutveri.fyi/mechanisms/deterministic-inference/evidence/flaws/2/) [12][16]. Emulating one GPU's rounding on another requires reverse-engineering tensor-core arithmetic and modelling proprietary kernel choices. Hawkeye covers a subset of NVIDIA architectures and states that attention and other higher-level operations need further reverse engineering. For the bit-exact emulator, a proprietary Hopper kernel family is an open edge case.

**Open minor failures**

- Mixed hardware widens the honest baseline (known failure, open question, in Sampled inference recomputation; https://trustbutveri.fyi/mechanisms/sampled-inference-recomputation/evidence/flaws/4/) [2]. 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.
- Deliberate faults leak a bit each (known failure, theoretical argument, in Low-trust AI compute verification system overview; https://trustbutveri.fyi/implementations/low-trust-compute-verification-system-overview/evidence/flaws/2/) [14]. A malicious device can leak one bit by deliberately outputting a wrong result, which blocks a disclosure when the cross-comparison fails. The design therefore needs a pre-agreed budget of tolerated faults.

**Scope limitations**

- Only recorded traffic is checked (scope limitation, theoretical argument, in Sampled inference recomputation; https://trustbutveri.fyi/mechanisms/sampled-inference-recomputation/evidence/flaws/2/) [1][13]. 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.

  Related mechanism: Network taps and certifiers (R1, not in the proposal). 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.
- Mismatches cannot be attributed to cheating or error (scope limitation, theoretical argument, in Low-trust AI compute verification system overview; https://trustbutveri.fyi/implementations/low-trust-compute-verification-system-overview/evidence/flaws/1/) [14]. A failed hash or replay does not show whether it came from an evasion attempt, a random bit flip or an evaluation error. The author notes that if detected anomalies can plausibly be waved off as malfunctions, deterrence becomes less effective, so the parties need an agreed escalation procedure that ends in attribution.

**Open questions**

- Inspector agents may be manipulable (open question, open question, in Low-trust AI compute verification system overview; https://trustbutveri.fyi/implementations/low-trust-compute-verification-system-overview/evidence/flaws/3/) [14]. Automated compliance screening with LLM-based inspector agents must resist prompt-injection attacks. Adversarially trained systems might hide malicious actions with steganography, which makes backdoor detection an open problem.

**Not yet demonstrated**

- Deterministic and bit-exact inference: Proposed (legacy code R1), assessed for screening challenged records to show declared inference compute is not training


## Possible additions

Mechanisms on the map, not in the proposal, that the records connect to an unaddressed or partly addressed claim, an open failure or a dependency. Pointers, not recommendations: each brings its own readiness level and findings, and none is claimed to close a failure.

- **Tamper evidence for verifier devices** (Research demonstration (legacy code R2), assessed for detecting probing of proposed verifier hardware, using server and electronics prototypes as evidence)
  - Deterministic and bit-exact inference waits on it: Tamper-evident, rapidly mass-manufacturable and retrofittable enclosures for side-channel defence are an open research question, and physical security against covert communication in every monitored data centre is challenging.
- **Timed challenge-response and memory-occupation challenges** (Research demonstration (legacy code R2), assessed for detecting whether a GPU is doing other work)
  - Deterministic and bit-exact inference waits on it: Distinguishing one server's DRAM contents from another's by challenge-response timing, and a general challenge-response protocol for diverse data types, are open.
- **Network taps and certifiers** (Proposed (legacy code R1), assessed for committing a complete record of cluster traffic, so declared inference can be checked)
  - 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.
  - Deterministic and bit-exact inference waits on it: Empirical feasibility of passive optical splitting at 53–112 GBaud under realistic conditions is an open question.
- **Side-channel suppression for isolated facilities** (Proposed (legacy code R1), assessed for bounding physical covert channels out of a verified enclosure)
  - Deterministic and bit-exact inference waits on it: A mass-manufacturable, good-enough side-channel defence, particularly power-line filtering, has not been constructed or red-teamed.


## Dependencies

**Missing prerequisites**

- Network taps and certifiers (Proposed (legacy code R1), assessed for committing a complete record of cluster traffic, so declared inference can be checked), needed by Deterministic and bit-exact inference

**Shared foundations**

- Network taps and certifiers, relied on by Sampled inference recomputation and Deterministic and bit-exact inference

**Blockers**

- 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) [8][17]
- Sampled inference recomputation: 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) [1][8]
- Sampled inference recomputation: 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) [8][9]
- Sampled inference recomputation: 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) [2][7]
- Sampled inference recomputation: 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) [2][13][14]
- Deterministic and bit-exact inference: Empirical feasibility of passive optical splitting at 53–112 GBaud under realistic conditions is an open question. (performance & compatibility; waits on Network taps and certifiers) [14]
- Deterministic and bit-exact inference: Exact replay needs complete hardware and software metadata, and the tolerable slowdown from emulation is an open question. (performance & compatibility; waits on Deterministic and bit-exact inference) [14]
- Deterministic and bit-exact inference: Tamper-evident, rapidly mass-manufacturable and retrofittable enclosures for side-channel defence are an open research question, and physical security against covert communication in every monitored data centre is challenging. (hardware trust; waits on Tamper evidence for verifier devices) [14]
- Deterministic and bit-exact inference: A mass-manufacturable, good-enough side-channel defence, particularly power-line filtering, has not been constructed or red-teamed. (coverage & hidden compute; waits on Side-channel suppression for isolated facilities) [14]
- Deterministic and bit-exact inference: Distinguishing one server's DRAM contents from another's by challenge-response timing, and a general challenge-response protocol for diverse data types, are open. (coverage & hidden compute; waits on Timed challenge-response and memory-occupation challenges) [14]
- Deterministic and bit-exact inference: The threat model is under-developed and needs input from cybersecurity and AI threat-modelling experts. (adversarial validation) [14]


## What the verifier sees

- Model weights: shown by none; depends on the design for Sampled inference recomputation; hidden by none; not involved in none; unspecified for Deterministic and bit-exact inference.
- Inputs and outputs: shown by none; depends on the design for Sampled inference recomputation; hidden by none; not involved in none; unspecified for Deterministic and bit-exact inference.
- Training data: shown by none; depends on the design for none; hidden by none; not involved in Sampled inference recomputation; unspecified for Deterministic and bit-exact inference.

## Implementations

- Sampled inference recomputation: [AI 2040 inference-only verification stack](https://trustbutveri.fyi/implementations/ai-2040-inference-only-verification-plan/) (R1, proposed architecture); [DiFR (Divergence From Reference)](https://trustbutveri.fyi/implementations/difr/) (R2, research prototype); [Low-trust AI compute verification system overview](https://trustbutveri.fyi/implementations/low-trust-compute-verification-system-overview/) (R1, proposed architecture); [SASH confidential network logger](https://trustbutveri.fyi/implementations/sash-confidential-network-logger/) (R1, research prototype); [TOPLOC](https://trustbutveri.fyi/implementations/toploc/) (R3, open-source project)
- Deterministic and bit-exact inference: [Batch-invariant inference kernels (Thinking Machines)](https://trustbutveri.fyi/implementations/batch-invariant-inference-kernels/) (R2, open-source project); [Verde and RepOps (Gensyn)](https://trustbutveri.fyi/implementations/gensyn-verde-repops/) (R3, product); [Low-trust AI compute verification system overview](https://trustbutveri.fyi/implementations/low-trust-compute-verification-system-overview/) (R1, proposed architecture)

## Sources

1. Verifying LLM Inference to Detect Model Weight Exfiltration, R. Rinberg et al. (2025). https://arxiv.org/abs/2511.02620
2. DiFR: Inference Verification Despite Nondeterminism, A. Karvonen et al. (2025). https://arxiv.org/abs/2511.20621
3. TOPLOC: A Locality Sensitive Hashing Scheme for Trustless Verifiable Inference, J. M. Ong et al. (2025). https://proceedings.mlr.press/v267/ong25a.html
4. PrimeIntellect-ai/toploc (GitHub repository), Prime Intellect (2025). https://github.com/PrimeIntellect-ai/toploc
5. INTELLECT-2: A Reasoning Model Trained Through Globally Decentralized Reinforcement Learning, Prime Intellect Team et al. (2025). https://arxiv.org/abs/2505.07291
6. adamkarvonen/difr (GitHub repository), A. Karvonen (2025). https://github.com/adamkarvonen/difr
7. Scaling Recomputation Inference Verification, Amodo Design (2026). https://amododesign.com/notes/2026-09-02-scaling-recomputation-inference-verification/
8. AI 2040 Plan A — Verification SITREP, Amodo Design (2026). https://amododesign.com/ai-verification/plan-a-sitrep/
9. Adversarial Entropy Inflation Against Gumbel-Based Inference Verification, N. Kezins (2026). https://arxiv.org/abs/2608.23375
10. SYNTHETIC-2 Release: Four Million Collaboratively Generated Reasoning Traces, Prime Intellect (2025). https://www.primeintellect.ai/blog/synthetic-2-release
11. An Inference Verification Prototype — Stage 1, Amodo Design (2026). https://amododesign.com/notes/2026-06-29-inference-verification-prototype/
12. Bit-Exact AI Inference Verification Without Performance Tradeoffs, N. Cankaya (2026). https://arxiv.org/abs/2606.00279
13. Example Schemes for Verifying High-Stakes AI Agreements, Amodo Design (2026). https://amododesign.com/notes/2026-06-23-verification-algorithms/
14. A System Overview for Near-Term, Low-Trust AI Compute Verification, N. Cankaya (2026). https://intelligence.org/wp-content/uploads/2026/06/A-system-overview-for-near-term-low-trust-AI-compute-verification.pdf
15. Fingerprinting All AI Cluster I/O Without Mutually Trusted Processors, N. Cankaya et al. (2026). https://arxiv.org/abs/2606.10724
16. Hawkeye: Reproducing GPU-Level Non-Determinism, E. Badash et al. (2026). https://proceedings.mlsys.org/paper_files/paper/2026/hash/e217c271a57c365a246b0ad39e668ba8-Abstract-Conference.html
17. Verification Plan, R. Dean (2026). https://ai-2040.com/supplements/verification-plan