Treatiesaim

Treatiesaim refers to a conceptual category of international instruments intended to govern artificial intelligence through treaty-based norms and obligations. The term appears in policy discussions to describe the goal of binding states and other actors to shared standards on AI safety, accountability, transparency, and ethics.

Although not a formal treaty itself, the idea emerged in the 2020s as scholars debated how governance principles could be translated into legally binding instruments that complement non-binding guidelines and industry codes of conduct.

A treatiesaim framework would typically address risks such as misalignment with human values, safety failures, dual-use concerns, and human rights impacts. Core design features include defined scope and parties, binding risk assessment and safety-by-design requirements, transparency and auditing obligations, independent oversight, and dispute-resolution mechanisms. Verification might involve international reporting and inspections, with sanctions for non-compliance and reinforced cooperation on research and information sharing.

None of the proposed instruments has been adopted to date. Critics warn that treatiesaims could hinder innovation, create implementation gaps, or lag behind rapid AI progress. Supporters contend that legally binding norms are necessary to manage cross-border risks and ensure accountability across actors.

Related concepts include global AI governance, arms-control-style regimes, export controls, and international human rights law. Notable proposed models discussed in policy circles include a Global AI Safety Treaty and a Treaty on Responsible AI Development.