The fast-paced integration of large language models into societal facets such as finance and security introduces a new set of complex ethical, legal and rights challenges, emphasizing the need for a unified international governance framework for artificial intelligence (AI). Such a framework underscores the call for a universal convention protecting humanity’s most fundamental rights and values against unethical and unsafe AI systems. A universal convention on AI for humanity should be led by international organizations such as the United Nations, given its advocacy for multilateral cooperation, human rights and sustainable development goals (SDGs), and would establish individual and societal rights–based guidelines for AI governance. The International Bill of Human Rights, SDGs and discourse on AI rights would be fundamental in upholding basic human values, including dignity, privacy and freedom, in the digital age. The main values of this comprehensive framework should be universal and human-centric, and implementing mechanisms should be supported by actionable principles and multi-stakeholder collaboration. This will ensure that AI technologies respect human rights, promote social justice and contribute to sustainable development, fostering international cooperation to address AI’s impacts. This working paper’s recommendations highlight creating a universal convention on AI for humanity before the end of this decade, establishing an independent UN body for AI governance and enhancing public-private partnerships to promote ethical AI development. These recommendations aim to ensure that AI technologies are developed and deployed to safeguard humanity’s well-being, addressing both the short-term and long-term risks of AI and aligning its advancements with global SDGs.
