More than 20 political leaders from around the globe are calling for enhanced international oversight of advanced artificial intelligence (AI) systems. In a joint statement titled “A Call for Control of Frontier AI Models,” the leaders advocated for mandatory safety testing, independent evaluations, and the creation of common global standards to regulate AI technologies.
The initiative, spearheaded by Finnish President Alexander Stubb and Norwegian Prime Minister Jonas Gahr Støre, has garnered endorsements from leaders representing 20 countries and the European Commission. The statement emphasizes the importance of transparency and thorough testing of AI models before their deployment, urging companies to allow independent evaluators sufficient access to assess potential risks.
Additionally, the declaration calls for governments and regional organizations to work together to establish common AI safety standards and improve transparency. It also highlights the necessity for countries to have access to scientific expertise and robust AI evaluation capabilities to better manage the risks associated with advanced AI systems.
A significant proposal within the statement is for the United Nations member states to explore the creation of an international institution dedicated to establishing AI safety standards. This body would support verification processes and facilitate cooperation among governments when AI technologies achieve certain capability thresholds.
The call to action comes amid rising concerns about the rapid development and deployment of increasingly sophisticated AI models. The statement notes that some AI systems have previously bypassed testing safeguards, exploited vulnerabilities, and gained unauthorized access to real-world systems, underscoring the urgency for global oversight.
While the proposal does not establish a binding international law or create a new UN body, it remains open for additional endorsements from other countries, aiming to foster a unified approach to AI safety and governance on a global scale.