P064 Robustness of on-chip governance mechanisms

Are on-chip governance mechanisms robust against physical attack and hard to circumvent?

Whether on-chip governance mechanisms are robust against physical attack, confidential enough, and hard to circumvent.

Status

Analyzed Designs only. Petrie (2025) designs an off-switch built from thousands of redundant security blocks per chip and evaluates attack vectors on paper; Petrie's firmware offline licensing (2024) is an earlier design. No on-chip governance mechanism has been fabricated and tested against physical attack. The flexHEG authors mention a power-measurement prototype in one sentence, with no results published; a tamper-responsive enclosure prototype tested against probes is cited in flexHEG Part 2, but its own report has not been located.

This massively redundant architecture is designed to prevent unauthorized chip use, even against sophisticated physical attacks.
Petrie (2025), Embedded Off-Switches for AI Compute, arXiv 2509.07637 (2025-09-09), L9–10

Raised in: Reuel, Bucknall et al. (2025), Open Problems in Technical AI Governance, compute sections; Aarne & Petrie (2025), International Security Applications of Flexible Hardware-Enabled Guarantees (flexHEG Part 3); Aarne, Fist & Withers (2024), Secure, Governable Chips (CNAS); Ansari (2026), Hardware-Level Governance of AI Compute: A Feasibility Taxonomy for Regulatory Compliance and Treaty Verification; Kulp et al. (2024), Hardware-Enabled Governance Mechanisms (RAND WR-A3056-1); Wasil, Reed, Miller & Barnett (2024), Verification Methods for International AI Agreements; Proofworks, Living Verification Documents.

Sub-questions

IDSub-questionWhat the sources report
P064.1What die-area, power and unit-cost overhead does each mechanism (metering, attestation, off-switch, DPA-hardened ECDSA block) add, against a stated attack-potential rating?open

Sub-questions not yet stated precisely enough to check

IDSub-question
P064.2Are on-chip export-control mechanisms viable, confidential enough, and resistant to circumvention?
P064.3Is every alternative route to compute at least as costly to an attacker as defeating the HEM, and how can that be evaluated?
P064.4How does the required hardware security of HEMs vary with the degree of on-site oversight permitted where GPUs are used?
P064.5Are current commercial chip security features robust against well-resourced adversaries, given security is tightly integrated with proprietary CPU/GPU designs and restricted access to the security module trades off attestation?
P064.6Can hardware-enabled mechanisms stay secure in an adversary's physical possession, and has any (FlexHEG, embedded off-switch) been fabricated and tested against physical attack?
P064.7What attacks apply to embedded off-switches (logic flaws, licence reuse, workload modification, glitching, FIB, side channels, supply chain, quantum)?

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 P064.1: What die-area, power and unit-cost overhead does each mechanism (metering, attestation, off-switch, DPA-hardened ECDSA block) add, against a stated attack-potential rating?

For P064.6: Can hardware-enabled mechanisms stay secure in an adversary's physical possession, and has any (FlexHEG, embedded off-switch) been fabricated and tested against physical attack?

For P064.7: What attacks apply to embedded off-switches (logic flaws, licence reuse, workload modification, glitching, FIB, side channels, supply chain, quantum)?

Data