Council of Europe AI Treaty Offers Governance Model Amid Disorder

A binding international treaty on artificial intelligence governance has emerged as a potential solution to regulatory fragmentation, even as the global order that enabled such agreements shows signs of strain.

The Council of Europe Framework Convention on AI, signed by 21 countries including all G7 members, represents the first international treaty requiring states to ensure AI systems respect human rights, democracy, and rule of law. The European Union ratified the agreement in May 2026, establishing a foundation that nearly a quarter of UN member states have now joined.

The governance gap

The treaty arrives as AI capabilities advance faster than regulatory frameworks can adapt. Recent disinformation campaigns demonstrate the technology's potential for manipulation at scale. Before Moldova's parliamentary elections, a Russian-funded network of just over 100 fake accounts generated 50 million views in under three months—in a country of 2.4 million people. Similar operations have targeted elections across Europe.

The challenge extends beyond electoral interference. AI systems now influence decisions about loans, employment, medical treatment, border control, and law enforcement. Without legal accountability, these systems can discriminate based on names, addresses, or demographic characteristics, with no mechanism for individuals to understand or challenge the outcomes.

Core principles for evolving technology

The Framework Convention takes a risk-based approach designed to remain relevant as AI capabilities evolve. Rather than attempting to regulate specific technologies, it establishes fundamental rights: people should know when machines are deciding their fate, understand the reasoning, and have the ability to challenge decisions.

This principles-based structure aims to address not just current AI systems but future developments, including autonomous agents that don't yet exist in widespread deployment.

Why it matters

Without coordinated international standards, AI governance risks fragmenting into competing regional frameworks and conflicting requirements. This fragmentation would create compliance complexity for developers, enable regulatory arbitrage, and leave populations in countries without strong governance frameworks vulnerable to exploitation. A common baseline helps prevent a race to the bottom while maintaining space for innovation.

The path forward

The treaty's open membership model distinguishes it from bloc-based approaches. Any nation can join, regardless of geography or existing alliances. This structure reflects recognition that AI governance cannot succeed as a tool of geopolitical competition.

The Framework Convention doesn't attempt to slow innovation or predict every AI application. Instead, it provides practical tools for risk assessment and accountability that countries can implement while adapting to local contexts.

These details were first reported in Time by a senior official involved in the treaty's development, writing ahead of the AI for Good Global Summit in Geneva.