# 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-0018,M-0001&implementations=M-0018:I-0009

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 |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Chip location verification / Lucid sovereignty (location) certificates | Proposed | Published security analysis | Semi-trusted | Analysis | Existing features | 0 / 2 / 0 | unspecified | unspecified | unspecified |
| Sampled inference recomputation | Operational use | Published security analysis | Adversarial | Analysis | None | 0 / 2 / 1 | depends | depends | not involved |

## Claims

No claims chosen.

## Mechanisms

### Chip location verification

A draft specification, hosted by Lucid Computing, for short-lived certificates that bound where a workload runs by timing signed messages to servers at known locations. ([Chip location verification](https://trustbutveri.fyi/mechanisms/chip-location-verification/))

- Assessment: selected implementation [Lucid sovereignty (location) certificates](https://trustbutveri.fyi/implementations/lucid-location-certificates/).
- Development: Proposed (legacy code R1), assessed for certifying the region where an attested workload ran at a given time.
- Security evidence: Published security analysis. Independent evaluation: unassessed. Formal proof: unassessed. Deployment assurance: unassessed.
- Claims in this proposal: none of them.
- Threat model: semi-trusted prover. Hardware: existing features. Prover cooperation: required. Attack testing: analysis. Category: Compute accounting & provenance.
- 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.

### 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.


## 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

**No new hardware needed**

- Chip location verification
- 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**

- Chip location verification / Lucid sovereignty (location) certificates: Analysis
- Sampled inference recomputation: Analysis


## Limits

**Open significant failures**

- On-chip keys may be extractable (known failure, theoretical argument, in Lucid sovereignty (location) certificates; https://trustbutveri.fyi/implementations/lucid-location-certificates/evidence/flaws/2/) [1][5]. Tee and Happel argue that ping-based location protocols backed by keys stored on the chip can be compromised if an adversary with physical access extracts those keys. In this specification, the evidence chain rests on the hardware root of trust's signed quote, whose signing key must be protected by the hardware.
- General delay and landmark attacks apply (known failure, theoretical argument, in Lucid sovereignty (location) certificates; https://trustbutveri.fyi/implementations/lucid-location-certificates/evidence/flaws/3/) [1][6]. Attacks on delay-based location verification in general also apply. Brass and Aarne discuss adding delay, using faster paths such as dark fibre, and compromising landmarks. The specification counters anchor impersonation with a signed anchor directory. Against collusion it recommends anchors in diverse places run by several independent operators, and peer monitoring that temporarily removes anchors whose timings deviate.
- 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/) [8][16][19]. 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 (R3, not 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/) [9][10]. 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 Lucid sovereignty (location) certificates. 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. Extracting a chip's key lets another device answer for it (known failure, theoretical argument, in Chip location verification; https://trustbutveri.fyi/mechanisms/chip-location-verification/evidence/flaws/1/) [5][6]. Ping-based protocols rely on cryptographic keys stored on the chip. Tee and Happel argue that an adversary with physical access could extract these keys and so compromise location verification. They propose GPU fingerprints as a mitigation, so far tested on 24 GPUs. Brass and Aarne assume the keys are stored securely, for example in a TPM.
- Context for Lucid sovereignty (location) certificates. 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. Added delay can shift an estimated position (known failure, demonstrated attack, in Chip location verification; https://trustbutveri.fyi/mechanisms/chip-location-verification/evidence/flaws/2/) [6][7]. Brass and Aarne cite internet-geolocation research in which artificially increased round-trip times moved the estimated location by up to 1,000 km, with a 74% chance of avoiding detection. Avellar and Grunewald list inflated ping times from circuitous routing as an evasion route. Added delay only loosens a distance bound, and Brass and Aarne propose a hard time limit as the counter: a chip that replies too slowly cannot be ruled out of a restricted location.
- Context for Lucid sovereignty (location) certificates. 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. Faster-than-assumed network paths (known failure, theoretical argument, in Chip location verification; https://trustbutveri.fyi/mechanisms/chip-location-verification/evidence/flaws/3/) [6][7]. Brass and Aarne list dark fibre and other private high-speed interconnects as ways to lower measured delays artificially. They judge that leasing dark fibre would probably not be a considerable challenge for covertly or openly adversarial actors. Avellar and Grunewald note that this can make a chip appear to be somewhere else entirely. A limit set at the vacuum speed of light cannot be beaten, but it makes honest chips fail more often.
- Context for Lucid sovereignty (location) certificates. 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. Compromised landmarks can falsify measurements (known failure, theoretical argument, in Chip location verification; https://trustbutveri.fyi/mechanisms/chip-location-verification/evidence/flaws/4/) [1][6][7]. A party that controls landmark servers can report false timing. Brass and Aarne cite research in which manipulating a third of the landmarks shifted the estimated location by about 700 km. Avellar and Grunewald note that compromised landmarks let adversaries spoof travel-time measurements directly. The draft specification asks verifiers to require anchors in diverse places, run by several independent operators.

**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/) [9]. 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.

**Scope limitations**

- Physical attacks on the trusted hardware are out of scope (scope limitation, theoretical argument, in Lucid sovereignty (location) certificates; https://trustbutveri.fyi/implementations/lucid-location-certificates/evidence/flaws/1/) [1]. The specification places the hardware root of trust and the TEE in the trusted computing base. It assumes they resist software attacks, notes that the attacker may have physical access, and leaves sophisticated physical attacks, such as bus probing and side-channel analysis, as a residual risk. It says that future revisions may add requirements for physical tamper evidence.
- Only recorded traffic is checked (scope limitation, theoretical argument, in Sampled inference recomputation; https://trustbutveri.fyi/mechanisms/sampled-inference-recomputation/evidence/flaws/2/) [8][20]. 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.

**Not yet demonstrated**

- Chip location verification: Proposed (legacy code R1), assessed for certifying the region where an attested workload ran at a given time


## 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.

- **Deterministic and bit-exact inference** (Operational use (legacy code R3), assessed for reproducing open-model inference from receipts in Gensyn's information-market service)
  - Bears on the open significant failure "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.
- **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.
- **TEE remote attestation for AI workloads** (Operational use (legacy code R3), assessed for showing which software ran to a party that distrusts the operator holding the hardware)
  - Chip location verification depends on it.


## Dependencies

**Missing prerequisites**

- TEE remote attestation for AI workloads (Operational use (legacy code R3), assessed for showing which software ran to a party that distrusts the operator holding the hardware), needed by Chip location verification

**Blockers**

- Chip location verification: The specification is an unfinished draft with no public implementation or evaluation. (adversarial validation) [1][2]
- Chip location verification: It needs a globally distributed, trusted anchor fleet and an endorser to run the anchor directory. (access & governance) [1]
- 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) [15][22]
- 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) [8][15]
- 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) [15][16]
- 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) [9][14]
- 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) [9][20][21]


## 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 Chip location verification.
- Inputs and outputs: shown by none; depends on the design for Sampled inference recomputation; hidden by none; not involved in none; unspecified for Chip location verification.
- Training data: shown by none; depends on the design for none; hidden by none; not involved in Sampled inference recomputation; unspecified for Chip location verification.

## Implementations

- Chip location verification: [Lucid sovereignty (location) certificates](https://trustbutveri.fyi/implementations/lucid-location-certificates/) (R1, standard)
- 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)

## Sources

1. Sovereignty Certificates: draft specification, version 0.1.0, Sovereignty Certificates Working Group (2025). https://github.com/Lucid-Computing/sovereignty-certificate-specification
2. Sovereignty Certificates Working Group (2026). https://sovcert.org/
3. Lucid Computing: Verifiable AI. Proven in hardware. (2026). https://lucidcomputing.ai/
4. Lucid Developer Platform documentation (2026). https://docs.lucidcomputing.ai/
5. GPU Fingerprinting for Location Verification, W. Tee & J. Happel (2026). https://arxiv.org/abs/2605.01930
6. Location Verification for AI Chips, A. Brass & O. Aarne (2024). https://www.iaps.ai/research/location-verification-for-ai-chips
7. Near-Term Verification Methods for AI Chip Exports, B. Avellar & E. Grunewald (2026). https://www.iaps.ai/research/near-term-verification-methods-for-ai-chip-exports
8. Verifying LLM Inference to Detect Model Weight Exfiltration, R. Rinberg et al. (2025). https://arxiv.org/abs/2511.02620
9. DiFR: Inference Verification Despite Nondeterminism, A. Karvonen et al. (2025). https://arxiv.org/abs/2511.20621
10. TOPLOC: A Locality Sensitive Hashing Scheme for Trustless Verifiable Inference, J. M. Ong et al. (2025). https://proceedings.mlr.press/v267/ong25a.html
11. PrimeIntellect-ai/toploc (GitHub repository), Prime Intellect (2025). https://github.com/PrimeIntellect-ai/toploc
12. INTELLECT-2: A Reasoning Model Trained Through Globally Decentralized Reinforcement Learning, Prime Intellect Team et al. (2025). https://arxiv.org/abs/2505.07291
13. adamkarvonen/difr (GitHub repository), A. Karvonen (2025). https://github.com/adamkarvonen/difr
14. Scaling Recomputation Inference Verification, Amodo Design (2026). https://amododesign.com/notes/2026-09-02-scaling-recomputation-inference-verification/
15. AI 2040 Plan A — Verification SITREP, Amodo Design (2026). https://amododesign.com/ai-verification/plan-a-sitrep/
16. Adversarial Entropy Inflation Against Gumbel-Based Inference Verification, N. Kezins (2026). https://arxiv.org/abs/2608.23375
17. SYNTHETIC-2 Release: Four Million Collaboratively Generated Reasoning Traces, Prime Intellect (2025). https://www.primeintellect.ai/blog/synthetic-2-release
18. An Inference Verification Prototype — Stage 1, Amodo Design (2026). https://amododesign.com/notes/2026-06-29-inference-verification-prototype/
19. Bit-Exact AI Inference Verification Without Performance Tradeoffs, N. Cankaya (2026). https://arxiv.org/abs/2606.00279
20. Example Schemes for Verifying High-Stakes AI Agreements, Amodo Design (2026). https://amododesign.com/notes/2026-06-23-verification-algorithms/
21. 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
22. Verification Plan, R. Dean (2026). https://ai-2040.com/supplements/verification-plan