Back to the MIT repository
6. Socioeconomic and Environmental3 - Other

Resistance to international law

AI models and systems may prove difficult to regulate or control under international law.

Source: MIT AI Risk Repositorymit1084

ENTITY

3 - Other

INTENT

3 - Other

TIMING

3 - Other

Risk ID

mit1084

Domain lineage

6. Socioeconomic and Environmental

262 mapped risks

6.5 > Governance failure

Mitigation strategy

1. **Develop and Ratify a Legally-Binding International Instrument** Establish a comprehensive international treaty or convention to mandate regulatory standards for general-purpose AI, prioritizing mechanisms that address systemic and existential risks. This instrument must incorporate provisions for regulatory convergence, drawing upon general principles of international law such as the precautionary principle, to ensure a unified and proactive global legal stance that is resilient against sovereignty challenges. 2. **Integrate Governance Disruption Resilience and Continuous Adaptation** Design the international regulatory framework with built-in mechanisms for continuous monitoring, evaluation, and legal adaptation. These mechanisms should ensure the framework remains responsive to rapid sociotechnical changes in AI capabilities, preventing governance development and destruction scenarios by maintaining legal certainty and institutional fitness for purpose. 3. **Codify Non-Derogable Human Rights and Control Principles** Embed explicit requirements into the international law that anchor AI development and deployment to fundamental human rights standards and the rule of law. Specifically, mandate accountability, transparency, and a principle of meaningful human control for AI systems that impact critical state functions or pose catastrophic risks, thus precluding AI from eroding the effectiveness and coherence of the global legal order.