Jun 14, 2025

Red Teaming Policy 5 of A Narrow Path: Evaluating the Threat Resilience of AI Licensing Regimes

Desmond, Wyatt Snyder

This report presents a red teaming analysis of Policy 5 from A Narrow Path, ControlAI’s proposal to delay Artificial Superintelligence (ASI) development through national AI licensing. Using a simplified PASTA threat modeling approach and comparative case studies (FDA, INCB, and California SB 1047), we identified two critical failure modes: regulatory capture and lack of whistleblower protections.

We developed a custom policy CVSS framework to assess cumulative risk exposure across each case. Due to time constraints, we used ChatGPT-assisted simulation to complete the results section and illustrate potential findings from our scoring method.

Our analysis suggests that, as written, Policy 5 is vulnerable to institutional influence and lacks sufficient safeguards to ensure enforcement. We recommend clearer accountability structures, built-in whistleblower protections, and stronger international coordination to make the policy more resilient.

Reviewer's Comments

Reviewer's Comments

Arrow
Arrow
Arrow
Arrow
Arrow

I think the possibility of regulatory capture of national regulators is a real concern and Gatling and Snyder are right to highlight it.

Their comparison tables seem useful.

I think their recommendations on accountability structures and whistleblower protections make sense. International coordination is sort of out of the scope of this policy, but I think it also makes sense to highlight it.

Appreciate the focus on specific failure modes, and the attempt to bring in an analytic framework. The emphasis on need to avoid regulatory capture and protect whistleblowers is fair. However, the discussion was overall fairly difficult to follow and lacked some of the argumentative depth about _why_ these were the most important considerations that need to be addressed that we were looking for.

Cite this work

@misc {

title={

(HckPrj) Red Teaming Policy 5 of A Narrow Path: Evaluating the Threat Resilience of AI Licensing Regimes

},

author={

Desmond, Wyatt Snyder

},

date={

6/14/25

},

organization={Apart Research},

note={Research submission to the research sprint hosted by Apart.},

howpublished={https://apartresearch.com}

}

Recent Projects

View All

View All

Feb 2, 2026

Prototyping an Embedded Off-Switch for AI Compute

This project prototypes an embedded off-switch for AI accelerators. The security block requires periodic cryptographic authorization to operate: the chip generates a nonce, an external authority signs it, and the chip verifies the signature before granting time-limited permission. Without valid authorization, outputs are gated to zero. The design was implemented in HardCaml and validated in simulation.

Read More

Feb 2, 2026

Fingerprinting All AI Cluster I/O Without Mutually Trusted Processors

We design and simulate a "border patrol" device for generating cryptographic evidence of data traffic entering and leaving an AI cluster, while eliminating the specific analog and steganographic side-channels that post-hoc verification can not close. The device eliminates the need for any mutually trusted logic, while still meeting the security needs of the prover and verifier.

Read More

Feb 2, 2026

Modelling the impact of verification in cross-border AI training projects

This paper develops a stylized game-theoretic model of cross-border AI training projects in which multiple states jointly train frontier models while retaining national control over compute resources. We focus on decentralized coordination regimes, where actors publicly pledge compute contributions but privately choose actual delivery, creating incentives to free-ride on a shared public good. To address this, the model introduces explicit verification mechanisms, represented as a continuous monitoring intensity that improves the precision of noisy signals about each actor's true compute contribution. Our findings suggest that policymakers designing international AI governance institutions face a commitment problem: half-measures in verification are counterproductive, and effective regimes require either accepting some free-riding or investing substantially in monitoring infrastructure.

Read More

This work was done during one weekend by research workshop participants and does not represent the work of Apart Research.
This work was done during one weekend by research workshop participants and does not represent the work of Apart Research.