P001 Seals and tamper-evident monitoring of servers and racks
What detection and false-accept rates do tamper-evident seals, tags and camera/sensor suites achieve on chips, servers and racks under skilled covert attack?
Detection rates of tamper-evident seals, tags and camera/sensor suites on chips, servers and racks (not tamper-respondent enclosures).
Status
Started Amodo's own status report of a built tray-level retrofit sensor suite (cameras with OpenCV intrusion detection, PIR, temperature, pressure, IMU, tamper mesh, lid switch, time-of-flight; STM32 TrustZone controller; power-down response) in an RFD still marked 'idea'; no detection or false-alarm rate, no skilled-attacker test (red-team and stress testing are listed as future work), no dataset. The seals clause has no held test: the 289-seal assessment that found all could be defeated cheaply is named by Aarne, Fist & Withers (2024) but not held, so it is not Failed here; 'TamperSec working on this (?)' in the Proofworks projects list is hearsay; Harack et al.'s glitter-and-photograph suggestion is untested. The headline's 'under skilled covert attack' clause has no work.
Amodo's v0.1 system includes 8 countermeasures selected from 38 candidates after reviewing 27 attack vectors. … Cameras (4 per tray, OpenCV-based intrusion detection) … Successfully demonstrated detection using STM built-in logic
Posed by
- Petrie & Aarne (2025), Technical Options for Flexible Hardware-Enabled Guarantees (flexHEG Part 2):
Some of the older tamper-evidence techniques may not be perfectly secure, like the cold-war seals that can apparently be broken
(Secure Enclosure, p19, L825–827) (source states the problem but not the quantity the question asks for)
Also raised in: Harack et al. (2025), Verification for International AI Governance (Oxford AIGI); Baker, Kulp, Marks, Brundage & Heim (2025), Verifying International Agreements on AI: Six Layers of Verification for Rules on Large-Scale AI Development and Deployment (RAND WR-A4077-1); Petrie, Aarne, Ammann & Dalrymple (2025), Overview of Flexible Hardware-Enabled Guarantees (flexHEG Part 1); Aarne, Fist & Withers (2024), Secure, Governable Chips (CNAS); Shavit (2023), What Does It Take to Catch a Chinchilla? Verifying Rules on Large-Scale Neural Network Training via Compute Monitoring; Hausenloy & Li (2026), The Compute Verification Post; AI Security Forum, Requests for Discussion; Amodo (2026), AI 2040 Plan A — How You Can Help; Dean, Verification Plan (AI 2040); Proofworks, Living Verification Documents.
Sub-questions
| ID | Sub-question | What the sources report |
|---|---|---|
| P001.1 | Is there a tag or seal that shows tampering, resists cloning, binds to a chip or board, and has a demonstrated false-accept rate (including low-tech glitter/photo seals)? | a source suggests an answer (not checked) |
| P001.2 | Does any tamper-evident seal survive an independent vulnerability assessment at a stated attacker budget, and what is the best seal's false-accept rate against a skilled covert attacker? | partly answered |
| P001.3 | What fraction of covert tamper attempts on seal-and-camera-protected (retrofit) AI servers are detected in a blind red-team, with a published false-accept rate? | implied by a source, not stated as open |
| P001.4 | In a trial, what fraction of staged tampering events (rack opened, device added) does a camera/acoustic/accelerometer/power suite detect, at what false-alarm rate? | partly answered |
| P001.5 | Is there a tamper-evident security camera robust to nanometre-scale attacks without excess false positives? | open |
| P001.6 | How many person-hours per rack does a tamper-evident retrofit (seals, wraps, cameras) of an existing data centre take? | implied by a source, not stated as open |
| P001.7 | What is the false-positive rate of an enclosure PUF under transport vibration and temperature extremes? | open |
Sub-questions not yet stated precisely enough to check
| ID | Sub-question |
|---|---|
| P001.8 | Can structured maintenance-access protocols prevent maintenance from being used to disguise tampering? |
| P001.9 | Does tamper evidence impose zero security cost on the prover? |
| P001.10 | How much information do cheap facility-level retrofits (meters, cameras, seals) give the verifier? |
Possibly relevant work found by automated search
Found by AI agents searching for this problem's sub-questions, mostly by keyword and citation in the OpenAlex scholarly database and on the web; a second AI pass reviewed each one, and we spot-checked some by hand. They seemed reasonably related, as far as one can tell from outside each paper's field. Some come from neighbouring fields, such as arms control and financial audit, and bear on the problem by analogy. Not reviewed against the status above; a paper listed here does not change that status. 928 of 1052 quotes checked word for word against the paper's text.
For P001.1: Is there a tag or seal that shows tampering, resists cloning, binds to a chip or board, and has a demonstrated false-accept rate (including low-tech glitter/photo seals)?
For P001.3: What fraction of covert tamper attempts on seal-and-camera-protected (retrofit) AI servers are detected in a blind red-team, with a published false-accept rate?
For P001.4: In a trial, what fraction of staged tampering events (rack opened, device added) does a camera/acoustic/accelerometer/power suite detect, at what false-alarm rate?
For P001.5: Is there a tamper-evident security camera robust to nanometre-scale attacks without excess false positives?
For P001.7: What is the false-positive rate of an enclosure PUF under transport vibration and temperature extremes?
* Full text not reviewed; more likely to contain errors.
Data
- problems.csv — one row per problem
- problem_status.csv — status, scope and evidence per problem
- problem_sources.csv — the posed-by quote and other documents per problem
- leaves.csv — all sub-questions, including those without a checkable form
- leaf_sources.csv — every source row behind each sub-question