Zhang Jiadong Yao Yani，Generative Artificial Intelligence Blame Avoidance and International Security: Generative Mechanism, Challenges, and Governance，International Review，No.4，2026

Abstract: The rise of generative artificial intelligence (GAI) is reshaping the boundaries of international security, yet the systemic problem of responsibility avoidance in its application has not received sufficient attention. This article argues that GAI blame avoidance refers to a pattern of behavior in which actors deliberately exploit the technological characteristics of GAI to evade, shift, or downplay their own responsibilities. This pattern manifests in three dimensions: diversification buck-passing, technologized attribution, and dilution of responsibility. This phenomenon is not driven by a single factor but is the result of the interplay among three mechanisms: technological particularities, cognitive biases of actors, and institutional lag. At the technological level, the deep dynamism of the black box, output uncertainty, and the polarization of the responsibility chain make causal tracing and singular attribution difficult. At the cognitive level, the myth of technological neutrality, excessive trust in and reliance on GAI, and the dilution of responsibility awareness drive actors to actively “de-responsibilize” their own conduct. At the institutional level, slow expansion of international legal norms, fragmentation of national standards, and the erosion of responsibility in the open-source ecosystem provide space for rule arbitrage. The interaction of the above three mechanisms brings a series of challenges to international security, including: the failure of accountability in combat decision‑making, which obstructs the attribution pathway under humanitarian law; the malfunction of cyberattack attribution judgment, which weakens deterrence mechanisms; the loss of responsibility locking in content manipulation, which triggers the dilemma of holding cognitive attacks accountable; and the instability of trust in rule consensus, which leads to the fragmentation of security governance. To efectively regulate this behavior, the international community should establish foundational governance principles. reconstruct the institutional architecture for responsibility allocation, set up international transparency and accountability mechanisms, and promote institutional coordination in global governance.

Key Words: Generative Artificial Intelligence; Blame Avoidance; International Security; Technological Attribution; Responsibility Allocation