2025人工智能在州政府中的应用：平衡创新、效率与风险研究报告

Balancing Innovation, Efficiency, and Risk AI in State Government: Balancing Innovation,Efficiency, and Risk Katherine Barrett and Richard GreeneExecutive Advisers, American Society for Table of Contents ForewordExecutive SummaryIntroductionControlling ExpectationsAI AdvantagesRoadblocksThe Risks Excitement among state leaders for the potential ofartificial intelligence is palpable; putting GenAI into AIIn StAtEGovErnmEnt: BAlAncInGInnovAtIon, EffIcIEncy,AndrISk www.businessofgovernment.org Foreword On behalf of the IBM Center for The Business of Government, weare pleased to present this new report,AI in State Government:Balancing Innovation, Efficiency, and Risk, by Katherine Barrett This timely report examines the rapid expansion of artificial intelligence (AI)applications across US states, with a particular focus on the acceleratingexperimentation with generative AI (GenAI)State governments are increas- Daniel J. ChenokExecutive DirectorIBM Center for Through an in-depth review of current pilot projects, emerging use cases, andearly implementation lessons, the authors offer a forward-looking perspectiveon how GenAI can serve as a collaborative partner for state employeesThereport maps areas where AI can complement, augment, or automate tasks Key recommendations include fostering cross-agency learning networks,investing in targeted workforce training and upskilling, and adoptinggovernance frameworks that balance innovation with responsible useBy This report extends the IBM Center’s commitment to exploring the intersectionof AI, workforce transformation, and government service deliveryIt builds onprior Center publications such asGenAI and the Future of Government Work,which explores the transformative potential of GenAI in reshaping theworkforce;Navigating Generative AI in Government, which outlines strategic Chris ShriverSenior Partner-VicePresident, SLED IndustryLeader, IBM We are grateful to the authors for this insightful contribution, and believethe report will serve as a valuable resource for state leaders seeking to Executive Summary "AI in State Government: Balancing Innovation, Efficiency, and Risk" delves into the rapidexpansion of artificial intelligence (AI) applications across U Sstate governments, with a specificfocus on the accelerating experimentation with generative AI (GenAI)The report highlights howstate governments are exploring GenAI to streamline operations, enhance service delivery, and Through an in-depth review of current pilot projects, emerging use cases, and implementationlessons, the report offers a forward-looking perspective on how GenAI can augment the work ofstate employeesAI can enhance, complement, or automate tasks within diverse state functions,from public health and transportation to education The analysis emphasizes the importance of The report reflects the importance of AI governance, recognizing that technology alone is notalways the answer Governance that includestransparency and public accountability can improvepublic trustVarious governance models are outlined, such as Ohio's structure requiring agencies The report also addresses the risks associated with GenAI, including hallucinations, biased data,and privacy and security concernsThe authors stress that while AI can augment workforce pro-ductivity, human judgment is necessary to address these and similar risks—like privacy issues Key recommendations include: •Invest in targeted workforce training:Ensure that government employees at all levels, aswell as the public, are adequately trained to understand the potential benefits and risks of AI •Adopt governance frameworks that drive innovation and responsible use:Implementevidence-based policies, use guidelines, and other measures to ensure effective and safe AI By following these strategies, states can cultivate a technologically empowered workforce that isresilient in an era of rapid digital change, while keeping the human element at the heart of Introduction The use of artificial intelligence has greatly expanded in the states, with the number of exper-imental uses of Generative AI (GenAI) growing every day“I think everybody is chasing AI oneway or the otherIt is a headliner for every conference,” said Alan Shark, co-chair of the “This is so transformationalThe impact of this is beyond what states have seen, and cer-tainly what CIOs have seen in the in the past,” said Doug Robinson, executive director of the Artificial intelligence is nothing particularly new, and the phrase AI is often used to cover allits variationsFor some time, it has had the capacity to analyze or classify data to help gov-ernments perform a variety of tasksIn recent years, GenAI emerged with the ability to create The next step, agentic AI, takes things a step furtherAgentic AI uses machine learningmodels to move toward an objective—like the elimination of unproductiveregulations—taking more autonomous actions, although human involvement IBm center for the Bus