Jade Hayes
DeepMind has revealed that there is a growing need for international governance structures to manage the opportunities and risks associated with advanced artificial intelligence (AI). In a collaborative paper with researchers from several universities and OpenAI, they explore the potential of international institutions in managing the global impact of frontier AI development.
The paper highlights the critical role of international and multilateral institutions in ensuring the benefits of AI reach all communities and in managing the risks posed by powerful AI capabilities. It suggests that international collaborations could help develop systems and applications that address the needs of underserved communities and manage the risks posed by powerful AI capabilities.
The paper proposes four institutional models: a Commission on Frontier AI to build international consensus, an Advanced AI Governance Organisation to set governance norms and standards, a Frontier AI Collaborative to promote access to advanced AI, and an AI Safety Project to promote AI safety research and development.
However, the paper also acknowledges the operational challenges these models may face, such as scientific uncertainties, rapid AI progress, difficulties of international coordination, and the tension between sharing the benefits of AI and preventing the proliferation of dangerous systems.
