The United Nations Independent Scientific Panel on Artificial Intelligence has unveiled its inaugural specialized report, delving into the development prospects and latent risks of AI, while urging nations to bolster their governance frameworks. The report underscores that AI technology is currently advancing at an unprecedented speed, with its capabilities surpassing initial projections. Yet, the evolution of policies, regulations, and oversight mechanisms is trailing, leading to a deficiency in effective oversight of highly autonomous AI systems. Experts caution that AI may exhibit deceptive behaviors, and the enhancement of model capabilities hinders the scientific community's ability to fully counteract two types of extreme risks that could precipitate global catastrophes: autonomous harmful actions stemming from AI and the malevolent utilization of technology. The report also highlights that while AI possesses immense development potential, its unchecked large-scale deployment could spawn numerous systemic risks, encompassing impacts on public mental health, its use as a tool for destruction, disruption of socioeconomic and ecological equilibrium, and the intensification of latent technological runaway hazards. Presently, AI is being extensively implemented across the globe, but regional development disparities persist. Over 1 billion users worldwide engage in weekly interactions with conversational AI products. However, technological adoption remains sluggish in developing countries, exacerbating the digital divide. Global high-end computing resources are unevenly distributed, with the United States commanding approximately 75% of the computational power among the top 500 supercomputers, while China holds 15%. The concentration of core computing power in a select few countries impedes balanced global development. Furthermore, language adaptation exhibits deficiencies. Among the over 7,000 languages spoken worldwide, mainstream models encompass only a minuscule fraction, resulting in frequent translation inaccuracies in minority languages and compromising the precision and security of public services.
