The US, UK, and EU have signed the world’s first legally binding international treaty on artificial intelligence. This historic agreement, made at a Council of Europe conference in Vilnius, aims to balance innovation with the protection of human rights and democratic values.
The United States, the European Union, and the United Kingdom have signed the first legally binding international treaty on artificial intelligence (AI). This landmark agreement was finalised at a Council of Europe conference in Vilnius, Lithuania, on Thursday.
The new treaty, officially named the 'Council of Europe Framework Convention on Artificial Intelligence, Human Rights, Democracy, and the Rule of Law,' is a significant step in the global effort to regulate AI. It aims to balance the rapid advancement of AI technologies with the need to protect human rights and democratic values.
This agreement is the result of years of negotiations among 57 countries, including major AI players like the US, UK, and EU, as well as other nations such as Japan, Canada, Israel, and Australia. The treaty seeks to address the risks associated with AI while promoting responsible innovation.
Marija Pejčinović Burić, Secretary-General of the Council of Europe, highlighted the treaty's importance, stating, “We must ensure that the rise of AI upholds our standards, rather than undermining them.” She described the treaty as an “open and inclusive” document with a “potentially global reach,” inviting more countries to sign and ratify it.
AI is increasingly integrated into various sectors, raising concerns about privacy, discrimination, and misuse. Shabana Mahmood, Britain’s justice minister, emphasised that the treaty is a crucial step in ensuring that new technologies do not compromise fundamental values like human rights and the rule of law.
The treaty requires signatories to be accountable for any harmful or discriminatory outcomes from AI systems. It mandates that AI systems respect privacy and equality rights and provides legal recourse for those affected by AI-related human rights violations. This framework allows countries to regulate AI effectively while fostering innovation.
Peter Kyle, the UK’s Minister for Science, Innovation, and Technology, noted that the treaty establishes a “baseline that goes beyond just individual territories,” emphasising its global significance.
In addition to the US, UK, and EU, several other countries signed the treaty, including Andorra, Georgia, Iceland, Norway, Moldova, San Marino, and Israel. Non-member states such as Argentina, Australia, Japan, and Mexico were involved in drafting the treaty and are expected to sign soon.
The broad international participation underscores the treaty's global importance, extending beyond the EU’s recent AI Act, which regulates AI within the bloc.
The treaty applies to both public and private-sector AI systems, requiring developers to align their systems with human rights obligations. It also ensures that democratic processes, like judicial independence, are not compromised by AI applications.
Despite its ambitious goals, the treaty has notable exemptions. AI systems used for national security or research and development are not subject to the same scrutiny as other sectors. This has raised concerns among civil society groups, including the European Center for Not-for-Profit Law (ECNL).
Francesca Fanucci, a legal expert with ECNL, criticised the treaty for being “watered down” with broad principles and caveats that may limit its enforceability. She expressed concerns about the treaty’s legal certainty and effectiveness.
While the treaty is seen as a crucial step towards global AI regulation, some critics argue that it lacks provisions for punitive measures like fines. Compliance will mainly be monitored through mechanisms that some believe may not be robust enough.
Still, the treaty is considered a significant first step in creating a unified global approach to AI regulation. Peter Kyle expressed optimism, stating that it is “the first agreement with real teeth globally,” bringing together diverse nations.
The treaty’s signing aligns with other global AI regulatory efforts, including the EU’s AI Act, which came into effect last month, the G7 AI pact from October 2023, and the Bletchley Declaration signed by 28 countries, including China and the US, in November 2023.