AI：政府路线图

The World Governments Summit is a global platform dedicated to shapingthe future of governments worldwide. Each year, the Summit sets the agendafor the next generation of governments with a focus on how they can harnessinnovation and technology to solve universal challenges facing humanity.The World Governments Summit is a knowledge exchange center atthe intersection of government, futurism, technology, and innovation.It functions as a thought leadership platform and networking hub forpolicymakers, experts and pioneers in human development.The Summit is a gateway to the future as it functions as the stage foranalysis of future trends, concerns, and opportunities facing humanity.It is also an arena to showcase innovations, best practice, and smartsolutions to inspire creativity to tackle these future challenges.To Inspireand EnableThe Next Generationof GovernmentsWorld Governments Summit2 Table of ContentsIntroductionThe AI RisksWhy Building An AI Roadmap Is So UrgentBalancing Risk And Reward: The Role Of GovernmentsThe AI OpportunityConclusion 061808241436 IntroductionArtificial intelligence (AI) is experiencinga period of explosive growth. However, thistransformative technology also presentssignificant challenges, raising concernsabout job displacement, bias, privacy, andenvironment impact.This report explores the current landscape ofAI, examining both its immense potential andthe critical risks that need to be addressed. Wedelve into the role of governments in managingthis powerful technology, focusing on strategiesfor harnessing its benefits while mitigating itsdownsides. A holistic approach with effectivepolicies and regulations is crucial to ensure AIbecomes a force for positive change, drivingeconomic growth, social progress, and a moresustainable future. Why BuildingAn AI RoadmapIs So UrgentSection 1 AI is at an inflection point.Today, more than ever, the technologyhas the capacity to transform nearlyevery aspect of life. This immensepotential is why governmentsworldwide are developing and debatingcomprehensive strategies to guide theresponsible deployment of AI tools thatprotect both businesses and consumerswhile fostering innovation.9 GPUs have since been utilized for the rapidmathematical calculations that are demandedby AI. Condor Galaxy 1 uses more than 72,000of these GPUs. While China and the US leadthe world in supercomputing capacity, not allgovernments have the same resources to deploysuch compute resources. Initiatives such asthe EU’s Joint Supercomputing Initiative helpalleviate such challenges, along with cloud-based supercomputing, although the latter isgenerally not used by governments given datasecurity concerns.Computational capacity of the fastest supercomputers in teraflops200020052010201520202025Source: Dongarra et al. (2023) — with minor processing by Our World in DataThis data is expressed in US$ per Terabyte (TB). It is adjusted for inflationMemoryFlashDiskSolid state1990200020102022Note: For each year, the time series shows the cheapest historical price recorded until that year.Source: John C. McCallum (2023); U.S. Bureau of Labor Statistics (2024) — with minor processing by Our World in Dataaccessible, while elastic cloud computing allowsfor flexibility in demand when it comes tocomputer processing, memory, and storage. While it may feel sudden, the rise of AI hasbeen occurring for decades, backed by adramatic increase in computing power. Forinstance, in November 2023, UAE-based G42and US-based Cerebras completed the CondorGalaxy 1 — a supercomputer four times morepowerful than the previous best.1There willbe nine such supercomputers combined intoan unprecedented single network by the endof 2024. Supercomputers are made possibleby graphics processing units (GPUs), originallydesigned for the video gaming industry toimprove computer graphics.Figure 100.1 billion0.3 billion0.5 billion0.7 billion0.9 billion1.1 billion19751980198519901995Figure 2Evolution of storage and memory cost100 trillion1 trillion100 billion100 million1 million10,000100195619701980Other contributing factors to AI’s growth includethe low cost of storage and computing power.Cloud storage is now ubiquitous and easily10 Additionally, advancements in the mathematicalunderpinnings of AI, especially machinelearning, or statistical learning, are pushing thetechnology forward. Researchers have developedmathematical techniques to convert complex,non-linear problems into linear formulationsFigure 3Research in AICount of publications, 1980-202160040020001980198519901995Machine learningComputer VisionArtificial neutral networkSource: White House, The Impact of AI on the Future of Workforces in the EU and USA, 2022And now companies of all types are funnelingbillions of dollars into generative AI, a sub-areaof AI that enables users to easily and quicklygenerate new content based on a variety of userprompts or inputs. The content generated bythese models includes text, images, sounds,animation, and 3D models (Figure