US, UK and EU sign first global treaty on AI use
The European Union, the United States and the United Kingdom have joined several other countries in signing what the Council of Europe describes as the first international legally binding treaty on artificial intelligence. The agreement, opened for signature in Vilnius, is meant to set a framework for how AI systems are developed and used across their full lifecycle.
The supplied DW report says the Council of Europe hailed the treaty as a milestone and said it was open to additional signatories. That matters because the treaty is being presented not as a narrow regional deal but as a document with the potential to grow into a wider international standard. The Council of Europe’s secretary-general said the goal was to ensure that AI advances do not undermine human rights, democracy or the rule of law.
The treaty’s supporters see it as a first attempt to put basic legal guardrails around a technology that is already changing public administration, law enforcement, business and digital communication. The treaty is said to cover the entire lifecycle of AI systems, which means it is intended to address not just deployment but also development and oversight. That breadth makes it an important marker in the global push to regulate AI before harms become harder to contain.
At the same time, the report also suggests the treaty is not as hard-edged as some critics wanted. Francesca Fanucci of the European Center for Not-for-Profit Law told Reuters, in comments carried by DW, that the convention had been watered down into broad principles and that some obligations may be too vague to enforce effectively. She also pointed to national security exemptions and relatively limited scrutiny of private companies compared with the public sector.
Those criticisms matter because they hint at the central tension in AI regulation: governments want rules that are strong enough to matter, but broad enough to win signatures from major powers. The fact that the US, UK and EU all signed the treaty is itself notable, given how differently those jurisdictions approach tech policy. It suggests at least a shared willingness to endorse common principles even if the details remain contested.
The treaty arrives after the EU already approved its own AI Act, which regulates high-risk uses of the technology. Together, the two moves show that AI governance is moving from abstract discussion toward formal legal architecture. Whether the treaty becomes a meaningful enforcement tool will depend on ratification, implementation and how aggressively signatories choose to interpret its standards.
For now, the signing is a political signal as much as a legal one. It shows that the world’s major democracies are trying to agree on the basic rules of the AI age before the technology outruns the law.
