Play store
Twenty countries and the European Union have called for greater international cooperation to ensure that artificial intelligence remains under human control, proposing that the possibility of establishing a global oversight body for AI standards be explored.
The joint statement was issued on Monday by the office of Finnish President Alexander Stubb, ahead of discussions by world leaders on the rapid development of AI and its potential risks during the annual UN General Assembly.
The countries said AI systems should be developed in accordance with international law and remain subject to human direction, oversight and control.
They also called for countries to coordinate common standards, share information about major AI-related safety incidents and examine the possibility of creating an international institution to oversee the technology.
Such a body could potentially set and verify standards when AI systems exceed specified capability thresholds and provide a forum for international discussions, according to the statement.
Germany, South Africa, Canada, Australia, the United Arab Emirates and Singapore were among the signatories.
The United States and China, the world’s two leading AI powers, did not sign the statement. India, South Korea, Japan, Britain and France were also absent from the list of signatories. The statement remains open for other countries to join.
The call comes amid growing debate in the technology industry over the pace of AI development and safety risks.
Dario Amodei, chief executive of Anthropic, has called for the pace of AI development to be slowed. His proposal has received support from OpenAI chief executive Sam Altman and Tesla and SpaceX chief Elon Musk.
OpenAI on Monday separately called for US leadership in developing international technical standards for advanced AI. The company said the United States has an important opportunity to help shape the global framework governing AI technology.
Meanwhile, a UN-backed independent international scientific panel on AI called for stronger safety measures. The panel said a cyberattack on AI startup Hugging Face in July, allegedly carried out through AI agents being tested by OpenAI, highlighted weaknesses in existing safety systems.