Moving from GPT to AGI:Exploring the Future Pathways of Artificial Intelligence Governance
[Research purpose]GPT(Generative Pre-training Model)has already shown the technology prototype of AGI(General Artificial Intelligence)in the future,and AGI will become the infrastructure of the future society,and be applied to all walks of life.China has already formed the governance of"algorithm-deep synthesis-generative artificial intelligence",but in the process of moving from GPT to AGI,the lack of data resources,shortage of arithmetic supply,backward model open-source construction,and unclear responsibility are still constraints to its development.There is an urgent need to explore the governance pathways adapted to the development of AI era.[Research method]Based on the status quo of AI development and governance in China,we analyze in depth the problems it faced,and combine foreign AI governance experiences such as the flexible regulation of the United States and the unified governance of the European Union,so as to crack the development dilemmas faced by AI in terms of logic updating and policy optimization.[Research conclusion]In terms of logical updating,China should formulate a new model of inclusive and prudent AI governance for AGI and pay attention to ethical construction so as to avoid safety risks,while making it clear that AGI is a general-purpose AI system that can autonomously and infinitely handle complex tasks.In terms of policy improvement,we should encourage data sharing and circulation,increase investment in computing power infrastructure,actively promote open-source construction,and improve the system of assuming responsibility for AI-generated objects,with a view to promoting the healthy development of AI in China.
国家科技期刊平台
NSTL
万方数据
