A New Social Risk Assessment Method for Deepfake Technology

Abstract

Abstract: As a new field of commercial application of artificial intelligence, deep synthesis is attracting more and more public attention. Deepfake technology is a destructive application of deep synthesis, which poses a great threat to national security, social security, and citizens' personal security. Its further development and evolution have been highly valued by governments, companies, and individuals in China, Europe and the United States. Given the lack of quantitative assessment methods and regulation system of a social risk of deepfake technology, this paper, based on the comprehensive analytic hierarchy process and fuzzy comprehensive evaluation, develops a new social risk assessment method to evaluate the social risk of deepfake technology. According to the evaluation results, this paper further puts forward the view that China's deepfake technology is currently in the stage of "medium and high" social risk and then reasonably infers that the main problem with China's Deepfake technology at present lies in:(1) immature theory and technology, (2) unsound legal supervision and (3) loopholes in products and services. The evaluation index system established in this paper can provide support for the application of deepfake technology and further social risk research in China, and the suggestions on regulating the social risk of deepfake technology can provide reference for the decision-making of relevant departments.

Key words: deepfake technology, emerging technology applications, network content security

Ma Lixia, Gao Gege, Wei Yunjie, He Ran. A New Social Risk Assessment Method for Deepfake Technology[J]. Management Review, 2024, 36(8): 119-132.

References

[1] 郭锐. 人工智能的伦理风险及其治理[J]. 北京航空航天大学学报(社会科学版), 2020,33(6):18-23 Guo R. Ethical Risk of AI and Its Governance[J]. Journal of Beijing University of Aeronautics and Astronautics(Social Sciences Edition), 2020,33(6):18-23
[2] Liu W., Hong Y. Research on the Challenge of Computer Artificial Intelligence Technology to Financial Risk Management[J]. Journal of Physics:Conference Series, 2021,1992(2):22-26
[3] 王砚羽,苏欣,谢伟. 商业模式采纳与融合:"人工智能+" 赋能下的零售企业多案例研究[J]. 管理评论, 2019,31(7):186-198 Wang Y. Y., Su X., Xie W. Adoption and Integration of Business Model in Retail Industry Empowered with "Artificial Intelligence +" -A Mult-case Study[J]. Management Review, 2019,31(7):186-198
[4] 卡鲁姆· 蔡斯著,张尧然,译. 人工智能革命——超级智能时代的人类命运[M]. 北京:机械工业出版社, 2017 Calum C. Surviving AI:The Promise and Peril of Artificial Intelligence[M]. Beijing:China Machine Press, 2017
[5] 何哲. 人工智能技术的社会风险与治理[J]. 电子政务, 2020,(9):2-14 He Z. Social Risk and Governance of Artificial Intelligence Technology[J]. E-Government, 2020,(9):2-14
[6] Bohannon J. Fears of an AI Pioneer[J]. Science, 2015,349(6245):252
[7] 马长山. 人工智能的社会风险及其法律规制[J]. 法律科学(西北政法大学学报), 2018,36(6):47-55 Ma C. S. Social Risk and Legal Regulation of Artificial Intelligence[J]. Science of Law(Journal of Northwest University of Political Science and Law), 2018,36(6):47-55
[8] Anderson J., Rainie L., Luchsinger A. Artificial Intelligence and the Future of Humans[J]. Pew Research Center, 2018,10(12):1-10
[9] 梁瑞刚,吕培卓,赵月,等. 视听觉深度伪造检测技术研究综述[J]. 信息安全学报, 2020,(3):1-17 Liang R. G., Lv P. Z., Zhao Y. A Survey of Audiovisual Deepfake Detection Techniques[J]. Journal of Cyber Security, 2020, (3):1-17
[10] Xia A. Chinese Deepfake APP Zao can Turn You into a Hollywood Actor-But should We Be Worried?[DB/OL]. https://www.thenationalnews.com/arts-culture/chinese-deepfake-app-zao-can-turn-you-into-a-hollywood-actor-but-should-we-be-worried-1.905747, 2019
[11] Zhavoronkov A., Ivanenkov Y. A., Aliper A., et al. Deep Learning Enables Rapid Identification of Potent DDR1 Kinase Inhibitors[J]. Nature Biotechnology, 2019,37(9):1038-1040
[12] Ivan F. Poacher Detection through Machine Learning[DB/OL]. https://silverpond.com.au/wp-content/uploads/2019/09/WPS-Casestudy-Highlighter-web.pdf, 2019
[13] 陈冬梅. 人工智能时代深度合成技术应用的风险及治理[J]. 辽宁行政学院学报, 2020,(5):76-79 Chen D. M. Risk and Governance of Deep Synthesis Technology Application in Artificial Intelligence Era[J]. Journal of Liaoning Academy of Governance, 2020,(5):76-79
[14] Leong D. Securing Elections and Beyond:Lessons for Singapore from Canada's 2019 Federal Election[R]. S. Rajaratnam School of International Studies, 2020
[15] 黎梦兵,吴勇. 新媒体的社会信任问题研究——基于"深度伪造"短视频视角[J]. 理论月刊, 2020,(12):81-90 Li M. B., Wu Y. Research on the Social Trust of New Media-Based on the Perspective of "Deeply Forged" Short Video[J]. Theory Monthly, 2020,(12):81-90
[16] Bateman J. Deepfakes and Synthetic Media in the Financial System:Assessing Threat Scenarios[R]. Carnegie Endowment for International Peace, 2020
[17] 王文娟,马方. "深度伪造"违法信息算法传播入罪的困境与破解[J]. 新闻界, 2021,(1):64-74 Wang W. J., Ma F. Difficulties and Solutions to the Crime of "Deep Fake" Illegal Information Algorithm Dissemination[J]. Journalism and Mass Communication, 2021,(1):64-74
[18] Smith H., Mansted K. Weaponised Deep Fakes:National Security and Democracy[R]. Australian Strategic Policy Institute, 2020
[19] Lossau N. Deep Fake:Gefahren, Herausforderungen und Lösungswege[R]. Konrad Adenauer Stiftung, 2020
[20] 纪晓东,薛晔,薛崇义. 区块链技术下互联网金融风险动态演变分析——基于AHP-FCM模型[J]. 企业经济, 2020,39(6):38-44 Ji X. D., Xue Y., Xue C. Y. Dynamic Evolution Analysis of Internet Financial Risk under Blockchain Technology-Based on AHP-FCM Model[J]. Enterprise Economy, 2020,39(6):38-44
[21] 迟国泰,迟枫,杜娟. 商业银行客户经理绩效评价体系与模型研究[J]. 管理评论, 2005,17(9):9-14 Chi G. T., Chi F., Du J. A Study on the Evaluation Systemand Model for Customer Managers' Performance in Commercial Banks[J]. Management Review, 2005,17(9):9-14
[22] 潘乾威,姚舜禹,陈玄默. 基于模糊层次分析的企业信贷信用风险评估模型研究[J]. 中国集体经济, 2021,(20):92-93 Pan Q. W., Yao S. Y., Chen X. M. Research on Credit Risk Assessment Model of Enterprise Credit Based on Fuzzy Hierarchical Analysis[J]. China Collective Economy, 2021,(20):92-93
[23] 金桂生,潘慧灵. 基于AHP的灰色关联法在企业战略决策中的应用研究[J]. 管理评论, 2011,23(9):137-143 Jin G. S., Pan H. L. The Applied Research of AHP-based Grey Interrelated Method on the Corporation's Strategic Decision-making[J]. Management Review, 2011,23(9):137-143
[24] Xu A., Zhang J., Li M. Comprehensive Evaluation Model of Livable City Based on Fuzzy Comprehensive Evaluation Method[J]. Asian Agricultural Research, 2018,10(4):65-69
[25] Yi K., Yu L., Lu L., et al. Research on the Weight Calculation of Social Benefit Evaluation of Chinese Film and TV Enterprises Based on Fuzzy Comprehensive Evaluation Method[J]. Journal of Intelligent & Fuzzy Systems, 2020,38(6):7813-7824
[26] 张立光,邱长溶. 社会保障综合评价指标体系和评价方法研究[J]. 管理评论, 2003,15(2):13-16 Zhang L. G., Qiu C. R. A Research on Establishing the Method and Index System for the Comprehensive Evaluation of China's Social Security[J]. Management Review, 2003,15(2):13-16
[27] Zhao F., Chen L. Empirical Study on the Teaching Quality Evaluation System of College Physics Informatization Based on Fuzzy Comprehensive Evaluation Method[J]. Revista de la Facultad de Ingenieria, 2017,32(14):58-66
[28] 林小荣. 基于灰色关联-层次分析的校园足球教学风险评价[J]. 吉林体育学院学报, 2020,36(3):63-71 Lin X. R. Risk Assessment of Campus Football Teaching Based on Grey Relational Hierarchy Analysis[J]. Journal of Jilin Sport University, 2020,36(3):63-71
[29] Wang G., Xiao C., Qi Z., et al. Development Tendency Analysis for the Water Resource Carrying Capacity Based on System Dynamics Model and the Improved Fuzzy Comprehensive Evaluation Method in the Changchun City, China[J]. Ecological Indicators, 2021,122:1-14
[30] 吕吉云. 军队传染病防控能力生成模式评估指标体系构建及应用研究[J]. 管理评论, 2015,27(5):114-120 Lv J. Y. Research on the Evaluation Criteria for Generating Pattern of PLA's Ability to Prevent and Control of Infectious Diseases[J]. Management Review, 2015,27(5):114-120
[31] 黄蕾,黄雨佳,刘朋辉,等. 区域综合环境风险评价方法体系研究[J]. 中国环境科学, 2020,40(12):5468-5474 Huang L., Huang Y. J., Liu P. H., et al. Research on Regional Comprehensive Environmental Risk Assessment Method System[J]. China Environmental Science, 2020,40(12):5468-5474
[32] 曹建峰,方龄曼. "深度伪造"的风险及对策研究[J]. 信息安全与通信保密, 2020,(2):89-97 Cao J. F., Fang L. M. Study on the Risk of "Deepfake" and the Related Countermeasures[J]. Information Security and Communications Privacy, 2020,(2):89-97