Leaders and representatives from 20 countries released a statement Monday afternoon calling for increased oversight of advanced AI systems.
The statement, posted by the president of Finland, declares that the rapid pace of AI development and recent episodes of AI systems’ sidestepping safety protocols demand concerted international coordination.
“AI must remain under human direction, oversight and control,” the statement reads. “It must be developed and used in line with international law.”
Representatives from the U.S., U.K. and China refrained from signing the document.
The statement comes a day before the United Nations General Assembly’s high-level general debate kicks off. During the annual event, political and business leaders from around the world descend on New York City for the week to hash out agreements and chart out priorities for the year ahead.
Signatories include Canadian Prime Minister Mark Carney, Kenyan President William Ruto, Netherlands Prime Minister Rob Jetten, Australian Prime Minister Anthony Albanese and South African President Cyril Ramaphosa.
Several ministers and other government figures signed the document as well, including Turkey’s Minister of Foreign Affairs Hakan Fidan, and the United Arab Emirates’ Deputy Prime Minister and Minister of Foreign Affairs Sheikh Abdullah bin Zayed Al Nahyan.
The statement features three main calls to action. First, it asks companies to test their systems for safety issues before they are publicly deployed, and allow independent evaluators to assess the risks from companies’ practices.
Second, the countries in the statement request that governments and regional groups develop common safety standards to ensure that countries around the globe can evaluate their models with required scientific expertise.
Third, the statement calls for U.N. members to leverage existing international efforts and “explore creating an international institution, able to set standards, enable verification, and convene states when capability thresholds are crossed.”
Nick Moës, executive director of AI governance nonprofit The Future Society, commended the initiative and said it was a sign of clear appetite for meaningful AI coordination. “It is great news that this broad range of governments come together to call for an international institution because no government can achieve this alone,” he told NBC News.
“But beyond statements, mobilizing diplomatic, political and financial capital will be needed to bring about the change that they agreed is necessary to keep control over frontier AI models,” Moes added.
At the General Assembly, AI will take center stage during official discussions and dozens of side events, with leaders aiming to set the AI agenda, secure investment and ensure their countries are relevant in AI debates. The U.N. Security Council and U.S. representatives will hold separate sessions dedicated to AI on Wednesday.
Several international efforts dedicated to AI and AI oversight already exist, but they vary in their remit and technological depth. The U.N. hosts the Independent International Scientific Panel on AI, a group of dozens of scientific experts from around the world, in addition to the Global Dialogue on AI Governance, a platform for countries to debate AI regulation and coordination.
