# 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-0012&implementations=M-0012:I-0022

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 |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Hardware-attested weight binding / PAL*M | Research demonstration | Published security analysis | Semi-trusted | Analysis | Existing features | 0 / 0 / 0 | unspecified | unspecified | unspecified |

## Claims

No claims chosen.

## Mechanisms

### Hardware-attested weight binding

A research prototype in which confidential-computing hardware signs hashes of the models, datasets and results involved in training, evaluation and inference. ([Hardware-attested weight binding](https://trustbutveri.fyi/mechanisms/model-identity-attestation/))

- Assessment: selected implementation [PAL*M](https://trustbutveri.fyi/implementations/palm/).
- Development: Research demonstration (legacy code R2), assessed for attesting declared model operations on a confidential CPU–GPU prototype.
- 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: On-chip & hardware-enabled.
- 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

**No new hardware needed**

- Hardware-attested weight binding


## 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**

- Hardware-attested weight binding / PAL*M: Analysis


## Limits

**Family finding context**

- Context for PAL*M. 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. Underlying attestation can be forged or relayed (known failure, demonstrated attack, in Hardware-attested weight binding; https://trustbutveri.fyi/mechanisms/model-identity-attestation/evidence/flaws/1/) [1][2][3][4][5][6]. Inherited finding. Critical for weight binding against an operator with physical access to affected hardware, or with control of the hypervisor on an AMD SEV-SNP platform without AMD's fixes. 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. Related finding: https://trustbutveri.fyi/mechanisms/tee-remote-attestation/evidence/flaws/1/.

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

  Related mechanism: Hardware-enabled guarantees (flexHEG) and guarantee processors (R1, not in the proposal). A tamper-protected enclosure around the chip is the proposed answer when the party that holds the hardware may attack it physically.
- Context for PAL*M. 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. Launch-state attestation does not by itself cover weights loaded later (known failure, theoretical argument, in Hardware-attested weight binding; https://trustbutveri.fyi/mechanisms/model-identity-attestation/evidence/flaws/2/) [7][8]. Attestation measures launch state, and weights are read from disk after boot. A signature checked at load time does not stop a malicious hypervisor from altering the disk afterwards. Tinfoil reports mitigating this with dm-verity checks on every read. Unmeasured runtime configuration remains a general risk.
- Context for PAL*M. 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. For private models, a user can confirm consistency but not content (scope limitation, open question, in Hardware-attested weight binding; https://trustbutveri.fyi/mechanisms/model-identity-attestation/evidence/flaws/3/) [7][9]. 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.

**Scope limitations**

- Physical and side-channel attacks are excluded (scope limitation, theoretical argument, in PAL*M; https://trustbutveri.fyi/implementations/palm/evidence/flaws/1/) [1]. The paper's guarantee assumes trusted TDX, CPU and H100 hardware. It treats side-channel defenses as orthogonal and excludes physical attacks, including memory-bus interposition and GPU replacement. These adversaries are outside the stated guarantee.


## 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.

- **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)
  - Hardware-attested weight binding 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 Hardware-attested weight binding

**Blockers**

- Hardware-attested weight binding: The paper plans prototype code release after peer review. (adversarial validation) [1]
- Hardware-attested weight binding: Overhead in memory-mapped dataset preprocessing and attribute-distribution experiments accounted for 37.64–42.50% of total instrumented runtime. (performance & compatibility) [1]


## What the verifier sees

- Model weights: shown by none; depends on the design for none; hidden by none; not involved in none; unspecified for Hardware-attested weight binding.
- Inputs and outputs: shown by none; depends on the design for none; hidden by none; not involved in none; unspecified for Hardware-attested weight binding.
- Training data: shown by none; depends on the design for none; hidden by none; not involved in none; unspecified for Hardware-attested weight binding.

## Implementations

- Hardware-attested weight binding: [Attestable Audits](https://trustbutveri.fyi/implementations/attestable-audits/) (R2, research prototype); [PAL*M](https://trustbutveri.fyi/implementations/palm/) (R2, research prototype); [Tinfoil model identity (Modelwrap)](https://trustbutveri.fyi/implementations/tinfoil-model-identity/) (R3, product)

## Sources

1. PAL*M: Property Attestation for Large Generative Models, P. Chantasantitam et al. (2026). https://arxiv.org/abs/2601.16199
2. TEE.fail: Breaking Trusted Execution Environments via DDR5 Memory Bus Interposition, J. Chuang et al. (2026). https://tee.fail/
3. Battering RAM: Low-Cost Interposer Attacks on Confidential Computing via Dynamic Memory Aliasing, J. De Meulemeester et al. (2026). https://batteringram.eu/
4. RMPocalypse: How a Catch-22 Breaks AMD SEV-SNP, B. Schlüter & S. Shinde (2025). https://rmpocalypse.github.io/
5. SEV-SNP RMP Initialization Vulnerability (AMD-SB-3020), AMD (2025). https://www.amd.com/en/resources/product-security/bulletin/amd-sb-3020.html
6. A primer on secure enclaves, Tinfoil (2026). https://docs.tinfoil.sh/verification/secure-enclave-primer
7. How Tinfoil Proves Exactly What Model Is Running, Tinfoil Team (2026). https://tinfoil.sh/blog/2026-02-03-proving-model-identity
8. On TEEs for Privacy-Preserving Monitoring in AI Governance, Gloria Z (2026). https://techgov.intelligence.org/blog/on-tees-for-privacy-preserving-monitoring-in-ai-governance
9. Attestable Audits: Verifiable AI Safety Benchmarks Using Trusted Execution Environments, C. Schnabl et al. (2025). https://arxiv.org/abs/2506.23706