About
Advanced artificial intelligence (AI) holds enormous promise, with the potential to accelerate scientific discovery, spur technological innovation and increase prosperity. At the same time, the risks associated with AI are growing, including issues of privacy, bias, information manipulation and inequality. Perhaps most critically, as AI systems become more powerful, they could pose safety risks worldwide. Among these global risks are the intentional misuse of powerful AI systems to cause widespread harm and the development of autonomous AI systems that cannot be controlled.
Successfully managing global AI risks requires breakthroughs in technical AI safety, such as improvements in the interpretability, controllability and alignment of AI systems. It also requires breakthroughs in governance: there is an urgent need for legitimate and effective decision making about how to develop and deploy advanced AI safely for all.
The Global AI Risks Initiative at the Centre for International Governance Innovation was created to advance the international governance that will be needed to manage global AI risks. The Initiative aims to mobilize the resources, talent and influence of policy makers, AI researchers, governance experts and civil society to reduce global risks from advanced AI systems. It will seek to build understanding of the importance of global risks from AI and access to workable policy options to mitigate these risks successfully.
The Global AI Risks Initiative will:
- work with partners to develop and test the necessary components of an international treaty/framework convention on advanced AI systems, including the principles, legitimate decision-making processes and implementation mechanisms required;
- build understanding among public policy makers and citizens of the strong scientific and technical case for global AI risks by providing tailored briefing materials and information sessions; and
- promote inclusive dialogue to strengthen global support for effective action on AI risks by convening consensus-building workshops across diverse perspectives.
