Selected work

Representative Papers

Verifying Restrictions on Frontier AI Research

Jun 26, 2026

This study addresses a critical challenge in AI governance: how to effectively verify compliance with restrictions on frontier artificial intelligence research amid insufficient international trust, thereby mitigating the existential risks posed by premature development of artificial superintelligence. The paper presents the first systematic framework for analyzing the verifiability of such research restrictions, integrating perspectives from policy, safety governance, and technical verification. It identifies and evaluates 28 candidate verification mechanisms—including training code audits, whistleblower protections, search warrants, and intelligence-gathering methods—assessing their feasibility and limitations. By establishing a comprehensive analytical foundation, this work fills a significant gap in the literature and provides both theoretical grounding and practical pathways for developing deployable verification tools to oversee advanced AI research.

0 citationsRead paper

Bit-Exact AI Inference Verification Without Performance Tradeoffs

May 29, 2026

This work addresses the challenge of achieving bit-level exact verification of AI inference, which is hindered by the non-determinism of GPU floating-point operations and forces existing auditing approaches to rely on approximate matching—rendering them vulnerable to stealthy attacks. The authors uncover the mechanisms by which modern inference engines (e.g., vLLM, Hugging Face Transformers) produce deterministic yet non-invariant outputs under default configurations. They propose a software-level recomputation method that enables bit-accurate verification across different NVIDIA GPU variants without requiring identical hardware or incurring performance overhead. By modeling floating-point behavior, reconstructing internal engine states, and analyzing deterministic execution paths, the approach transforms accumulated rounding errors into auditable software-hardware fingerprints, establishing a tamper-resistant and verifiable foundation for AI governance.

0 citationsRead paper

Does Distributed Training Undermine Compute Governance?

May 28, 2026

This study addresses the challenge that cutting-edge AI model training can circumvent existing computational governance by leveraging distributed hardware, thereby undermining regulatory effectiveness. It presents the first systematic analysis of how distributed training complicates compute governance and proposes a multidimensional countermeasure framework integrating algorithmic characteristics, hardware-level tracking, cluster-wide compute thresholds, whistleblower mechanisms, and legal audits. The research delineates the technical boundaries of evasion tactics and advances practical regulatory strategies, including chip-level monitoring and memory/compute usage thresholds. These contributions offer both theoretical grounding and actionable pathways for strengthening the governance of AI compute infrastructure.

0 citationsRead paper

An International Agreement to Prevent the Premature Creation of Artificial Superintelligence

Nov 13, 2025

This paper addresses extinction-level risks—such as objective misalignment, geopolitical conflict, and malicious misuse—posed by premature development of Artificial Superintelligence (ASI). Methodologically, it proposes a U.S.–China–centered international governance framework featuring a novel verifiable technical architecture that integrates FLOP-based training thresholds, end-to-end AI chip lifecycle tracking, and model usage behavior attestation—embedded within legally binding prohibitions and multilateral negotiation mechanisms to balance security and feasibility. Its primary contribution is the first ASI development constraint pathway that is technically enforceable, independently verifiable, and politically acceptable amid profound trust deficits—effectively halting high-risk capability jumps while preserving beneficial AI applications. The framework offers an operationally viable policy paradigm for global AI safety governance; however, its implementation hinges on sustained great-power political consensus and adaptive alignment with technological evolution.

0 citationsRead paper

Governing AI R&D: A Legal Framework for Constraining Dangerous AI

Sep 03, 2025

This study addresses public safety risks arising from AI development and examines constitutional challenges to high-risk AI regulation in the U.S. context. Methodologically, it systematically analyzes the U.S. constitutional and administrative law frameworks—focusing on First Amendment protections (speech and commercial expression), administrative law constraints (delegation doctrine and procedural legitimacy), and Fourteenth Amendment guarantees (due process and equal protection). Integrating case-law analysis, constitutional interpretation, and policy compliance assessment, the study constructs the first comprehensive litigation risk map for AI regulation and proposes a precedent-informed, preventive legislative framework. Its contributions include: (1) delineating the constitutional boundaries of AI R&D oversight; (2) specifying institutionally viable, judicially defensible regulatory designs; and (3) demonstrating that rigorous, constitutionally sound AI governance is legally feasible. The findings offer both theoretical foundations and actionable policy blueprints for global AI法治ization (rule-of-law-based AI governance).

0 citationsRead paper

Recent publications

Latest Papers

Verifying Restrictions on Frontier AI Research

Jun 26, 2026

This study addresses a critical challenge in AI governance: how to effectively verify compliance with restrictions on frontier artificial intelligence research amid insufficient international trust, thereby mitigating the existential risks posed by premature development of artificial superintelligence. The paper presents the first systematic framework for analyzing the verifiability of such research restrictions, integrating perspectives from policy, safety governance, and technical verification. It identifies and evaluates 28 candidate verification mechanisms—including training code audits, whistleblower protections, search warrants, and intelligence-gathering methods—assessing their feasibility and limitations. By establishing a comprehensive analytical foundation, this work fills a significant gap in the literature and provides both theoretical grounding and practical pathways for developing deployable verification tools to oversee advanced AI research.

0 citationsRead paper

Bit-Exact AI Inference Verification Without Performance Tradeoffs

May 29, 2026

This work addresses the challenge of achieving bit-level exact verification of AI inference, which is hindered by the non-determinism of GPU floating-point operations and forces existing auditing approaches to rely on approximate matching—rendering them vulnerable to stealthy attacks. The authors uncover the mechanisms by which modern inference engines (e.g., vLLM, Hugging Face Transformers) produce deterministic yet non-invariant outputs under default configurations. They propose a software-level recomputation method that enables bit-accurate verification across different NVIDIA GPU variants without requiring identical hardware or incurring performance overhead. By modeling floating-point behavior, reconstructing internal engine states, and analyzing deterministic execution paths, the approach transforms accumulated rounding errors into auditable software-hardware fingerprints, establishing a tamper-resistant and verifiable foundation for AI governance.

0 citationsRead paper

Does Distributed Training Undermine Compute Governance?

May 28, 2026

This study addresses the challenge that cutting-edge AI model training can circumvent existing computational governance by leveraging distributed hardware, thereby undermining regulatory effectiveness. It presents the first systematic analysis of how distributed training complicates compute governance and proposes a multidimensional countermeasure framework integrating algorithmic characteristics, hardware-level tracking, cluster-wide compute thresholds, whistleblower mechanisms, and legal audits. The research delineates the technical boundaries of evasion tactics and advances practical regulatory strategies, including chip-level monitoring and memory/compute usage thresholds. These contributions offer both theoretical grounding and actionable pathways for strengthening the governance of AI compute infrastructure.

0 citationsRead paper

An International Agreement to Prevent the Premature Creation of Artificial Superintelligence

Nov 13, 2025

This paper addresses extinction-level risks—such as objective misalignment, geopolitical conflict, and malicious misuse—posed by premature development of Artificial Superintelligence (ASI). Methodologically, it proposes a U.S.–China–centered international governance framework featuring a novel verifiable technical architecture that integrates FLOP-based training thresholds, end-to-end AI chip lifecycle tracking, and model usage behavior attestation—embedded within legally binding prohibitions and multilateral negotiation mechanisms to balance security and feasibility. Its primary contribution is the first ASI development constraint pathway that is technically enforceable, independently verifiable, and politically acceptable amid profound trust deficits—effectively halting high-risk capability jumps while preserving beneficial AI applications. The framework offers an operationally viable policy paradigm for global AI safety governance; however, its implementation hinges on sustained great-power political consensus and adaptive alignment with technological evolution.

0 citationsRead paper

Governing AI R&D: A Legal Framework for Constraining Dangerous AI

Sep 03, 2025

This study addresses public safety risks arising from AI development and examines constitutional challenges to high-risk AI regulation in the U.S. context. Methodologically, it systematically analyzes the U.S. constitutional and administrative law frameworks—focusing on First Amendment protections (speech and commercial expression), administrative law constraints (delegation doctrine and procedural legitimacy), and Fourteenth Amendment guarantees (due process and equal protection). Integrating case-law analysis, constitutional interpretation, and policy compliance assessment, the study constructs the first comprehensive litigation risk map for AI regulation and proposes a precedent-informed, preventive legislative framework. Its contributions include: (1) delineating the constitutional boundaries of AI R&D oversight; (2) specifying institutionally viable, judicially defensible regulatory designs; and (3) demonstrating that rigorous, constitutionally sound AI governance is legally feasible. The findings offer both theoretical foundations and actionable policy blueprints for global AI法治ization (rule-of-law-based AI governance).

0 citationsRead paper