当前位置:首页 > 报告详情

AI数据以及国民帐户系统的更新.pdf

上传人: 鲁** 编号:615347 2025-03-03 9页 685.11KB

1、IMF IMF|Statistics1AI,Data,and the Update of the System of National AccountsNOVEMBER 21,2024Erich StrassnerChief,Real Sector Statistics DivisionIMF Statistics DepartmentIMF IMF|Statistics2AI and Data:MotivationAI and Data:MotivationSince the 2008 System of National Accounts(SNA)was published,product

2、ion of data is now fundamental to many business models,while most traditional businesses now produce some form of data to either increase returns or lower costs.A large amount of this data is used repeatedly in production for more than one year,but no fixed asset category currently exists outside of

3、 possible inclusion within databases.In addition,Generative AI has been progressively developing over the years,but a significant leap in usability and public awareness occurred around 2020 with the release of models like OpenAIs GPT-3.This model demonstrated advanced capabilities in natural languag

4、e processing,enabling a wide range of applications in various industries.Moreover,data is an input into the development of Large Language Models(LLMs).If we do not value data as an asset we may be mis-representing the LLM production function.Because of the current classification and definition withi

5、n the SNA,there has been growing concern that data and AI are not visible and being appropriately represented in the accounts,impacting the estimates of GDP,balance sheets,and productivity.Source:CIOSource:PixabyIMF IMF|Statistics3Investment in Intangible Assets(1)Investment in Intangible Assets(1)S

6、ource:World Intangible Investment Highlights:https:/www.wipo.int/edocs/pubdocs/en/wipo-pub-rn2024-32-en-world-intangible-investment-highlights.pdf IMF IMF|Statistics4Investment in Intangible Assets(2)Investment in Intangible Assets(2)Source:World Intangible Investment Highlights:https:/www.wipo.int/

word格式文档无特别注明外均可编辑修改,预览文件经过压缩,下载原文更清晰!
三个皮匠报告文库所有资源均是客户上传分享,仅供网友学习交流,未经上传用户书面授权,请勿作商用。
本文讨论了人工智能(AI)和数据在国民经济核算中的地位和作用。自2008年国民账户体系(SNA)发布以来,数据生产已成为许多商业模型的基础,传统企业也通过数据提高回报或降低成本。文章指出,大量数据在生产中重复使用超过一年,但现有SNA体系中缺乏固定的资产分类来充分体现这一点。2020年OpenAI的GPT-3发布后,AI的可用性和公众认识有了显著提升,其在自然语言处理方面的能力推动了各行各业的应用。数据作为大型语言模型(LLM)开发的输入,如果不将其视为资产,可能会导致对LLM生产功能的错误表述。SNA当前的分类和定义未能充分反映数据和AI在账户中的可见性和代表性,影响GDP、资产负债表和生产力的估算。为此,文章提出了包括更新SNA中知识产权产品的定义,为AI提供明确的可视性,并将其添加到计算机软件下的“其中”分类中,以及定义数据作为生产结果,并将其分类为一种新资产类别“数据和数据库”的建议。此外,建议创建一个固定资产的新分类“计算机软件、数据和数据库”,并对AI和数据资产进行估值。目前, Eurostat-IMF任务小组正在开发一份关于测量数据作为资产的手册,并把估算AI软件在国民经济中的价值作为下一步优先任务。
"AI和数据如何影响国民经济统计体系?" "人工智能和数据资产如何被纳入新的国民经济分类?" "如何评估AI和数据资产的价值?"
客服
商务合作
小程序
服务号
折叠