The Council of Europe’s international AI treaty was signed today (September 5) in Lithuania by countries including the EU, Norway, the UK, and the US. This treaty, adopted in May, is the first international legally binding agreement that governs the entire lifecycle of AI systems. It aims to mitigate risks while encouraging responsible innovation.
Unlike the EU’s AI Act, which took effect in early August, the treaty allows non-EU countries to join. It also applies to both public and private sectors, offering different approaches for regulating AI in the private sector.
Council of Europe Secretary-General Marija Pejčinović Burić expressed hope that more countries will sign and ratify the treaty quickly so it can take effect soon.
The treaty will come into force three months after at least five countries, including three Council of Europe members, have ratified it. A total of 46 member states, including non-EU countries like Australia, Canada, Israel, Japan, and Argentina, participated in negotiating the treaty, which is open to countries worldwide.
Meanwhile, other regions are also advancing AI initiatives. The AI Safety Summit aims to further global discussions on AI safety, while Japan’s G7 Hiroshima process promotes secure and trustworthy AI through guiding principles and a voluntary Code of Conduct, complementing the EU’s AI Act.