| 176 | 0 | 33 |
| 下载次数 | 被引频次 | 阅读次数 |
人工智能技术的高速发展引领了人类科技的划时代变革，但算法模型、生产要素的高度集中也产生了垄断隐忧。人工智能技术运用可以细分为开发层面、模型层面以及应用层面，就开发层面而言，市场结构的失衡和数据垄断问题可能导致人工智能供应链中严重的竞争问题。人工智能基础模型的开发在硬件、云服务、训练数据等方面的需求，导致上游大型供应商与人工智能模型开发商的议价能力存在天然不对等，优质数据垄断的现实问题使得初创企业难以研发出与大型科技公司性能相当的基础模型，削弱了竞争效应。为此，应当参照必需设施原则，合理、有序地开放认定为必需设施的优质数据集，同时避免对促进数字生态系统竞争的非横向整合施加不合理阻碍，增强数字生态系统之间的竞争强度。
Abstract：The rapid advancement of artificial intelligence(AI) technology has ushered in an epoch-making transformation in human science and technology. Nevertheless, the high concentration of algorithm models and production factors has also given rise to concerns regarding monopolization. The operation of AI technology can be classified into the development layer, the model layer, and the application layer. At the development layer, the imbalance in the market structure and issues related to data monopolies might result in severe competition problems within the AI supply chain. The demanding requirements on hardware, cloud services, and training data, etc., in developing basic AI models, lead to an inherent imbalance in bargaining power between large upstream suppliers and AI model developers. The reality of monopolized high-quality data makes it challenging for start-ups to develop basic models with performance comparable to those of large technology companies, thereby weakening their competitive effect. Hence, it is necessary to refer to the essential facility doctrine, reasonably and orderly open high-quality datasets identified as essential facilities, and simultaneously avoid imposing unreasonable hindrances on non-horizontal integrations that promote competition in the digital ecosystem, enhancing the intensity of competition among ecosystems.
[1] 丁晓东.论算法的法律规制[J].中国社会科学，2020(12):138-159+203.
[2] Colorado General Assembly.Consumer protections for artificial intelligence[EB/OL].[2024-03-17].https://leg.colorado.gov/bills/sb24-205.
[3] Competition and Markets Authority.AI foundation models initial report[EB/OL].[2023-09-18].https://www.gov.uk/government/publications/ai-foundation-models-initial-report?gh_src=v3scng1.
[4] SCHNEIDER J,MESKE C,KUSS P.Foundation models[J].Business and information systems engineering,2024(2):221-231.
[5] 吴再龙，王利明，徐震，等.GPU虚拟化技术及其安全问题综述[J].信息安全学报，2022(2):30-58.
[6] 李希梁，张钦昱.生成式人工智能的反垄断规制[J].电子政务，2024(5):53-63.
[7] 张欣.面向产业链的治理：人工智能生成内容的技术机理与治理逻辑[J].行政法学研究，2023(6):43-60.
[8] 梅夏英，王剑.“数据垄断”命题真伪争议的理论回应[J].法学论坛，2021(5):94-103.
[9] 徐文.数字平台企业数据流通的竞争法检视及完善[J].法律科学(西北政法大学学报),2025(6):118-128.
[10] 王健，吴宗泽.论数据作为反垄断法中的必要设施[J].法治研究，2021(2):102-114.
[11] 孙晋.数字平台的反垄断监管[J].中国社会科学，2021(5):101-127+206-207.
[12] 许可.数据爬取的正当性及其边界[J].中国法学，2021(2):166-188.
[13] Competition and Markets Authority.Microsoft Corporation’s partnership with OpenAI,Inc.Decision on relevant merger situation SUMMARY[EB/OL].[2025-03-05].https://assets.publishing.service.gov.uk/media/67c841d6d0fba2f1334cf 276/1._MS.OAI_-_Summary.pdf.
[14] 周围.规制平台封禁行为的反垄断法分析——基于自我优待的视角[J].法学，2022(7):163-178.
[15] Competition and Markets Authority.Cloud infrastructure services provisional decision report[EB/OL].[2025-01-28].https://assets.publishing.service.gov.uk/media/6798ecb1419bdbc8514fdf8a/cloud_mi_provisional_decision_report.pdf.
[16] 王健，吴宗泽.生成式人工智能反垄断论纲[J].法治研究，2024(6):130-147.
[17] 王晓晔.反垄断新思潮新布兰代斯主义批判[J].法学评论，2025(1):26-38.
[18] European Commission.Commission clears acquisition of ZT by AMD[EB/OL].[2025-03-13].https://ec.europa.eu/commission/presscorner/detail/en/mex_25_766.
[19] 国家市场监督管理总局.英伟达涉嫌违反反垄断法市场监管总局依法决定立案调查[EB/OL].[2024-12-09].https://www.samr.gov.cn/xw/zj/art/2024/art_ed4d3090401741a0894e475d35db652b.html.
[20] 裴轶，马婷.数据开放与权力边界：公共数据的行政垄断风险识别与规制[J].中国信息安全，2025(1):23-28.
[21] GAMBARO M.Big data competition and market power[J].Market and competition law review,2018(2):99-122.
[22] 杨东.论反垄断法的重构：应对数字经济的挑战[J].中国法学，2020(3):206-222.
[23] 裴轶，韦武.企业向公共部门共享数据的欧盟模式与中国镜鉴——以欧盟《数据法案》为视角[J].南海法学，2024(3):86-101.
[24] JACOBIDES M G,CENNAMO C,GAWER A.Towards a theory of ecosystems[J].Strategic management journal,2018(8):2255-2276.
[25] 丁晓东.论数据垄断：大数据视野下反垄断的法理思考[J].东方法学，2021(3):108-123.
[26] HOVENKAMP H.Antitrust and platform monopoly[J].Yale law journal,2021(8):1952-2050.
[27] 李兆阳.从数据垄断走向数据开放：数据成为必需设施的竞争法分析[J].重庆大学学报(社会科学版),2023(6):190-204.
[28] 宁度.拒绝交易的反垄断法规制——兼评《关于平台经济领域的反垄断指南》拒绝交易条款[J].法治研究，2021(4):148-160.
[29] 孟雁北.数字经济时代反垄断法视野下必需设施理论之高度审慎适用[J].中国社会科学院大学学报，2024(11):41-57+160-161+165.
[30] 王晓晔.标准必要专利反垄断诉讼问题研究[J].中国法学，2015(6):217-238.
(1)英伟达公司收购迈络思科技有限公司股权案,市场监管总局公告﹝2020﹞第16号。
(2)See Merger Guidelines 2023,Part.3,Section 3.3.
(3)See United States v.Terminal Railroad Association,224 U.S.383(1912).
(4)See MCI Communications Corp.v.American Tel.&Tel.Co.,708 F.2d 1081(1983).
(5)See IMS Health GmbH & Co.OHG v.NDC Health GmbH & Co.KG,Case C-418/01[2004]ECRI-05039.
(6)See Gesetz gegen Wettbewerbsbeschr?nkungen,Part.1,Section 19(2).
(1)See Regulation(EU)2024/1689, Recital 1.
基本信息:
DOI：10.16452/j.cnki.sdkjsk.2025.06.007
中图分类号:D922.294
引用信息:
[1]裴轶,汪雨波.人工智能供应链中的垄断风险与竞争法因应[J].山东科技大学学报(社会科学版),2025,27(06):35-43.DOI:10.16452/j.cnki.sdkjsk.2025.06.007.
基金信息:
2025年度工信部指导性软课题(GXZK2025-2-57)