收藏切换
Security of generative artificial intelligence: Challenges, countermeasures, and future
收藏切换
PDF
Gansen ZHAO1, 4, Wenfeng XU1, 4, Cheng QIAN1, 4, Zhihao HOU1, 4, Mengqin NING2, Yue GONG2, Guangyuan KONG1, 4, Xiangmin XU5, 6, Jiahong GUO2, *, Wenjun MA1, 3, *
Science & Technology Review | 2026, 44(12) : 68 - 86
Less
收藏切换
Science & Technology Review | 2026, 44(12): 68-86
Security of generative artificial intelligence: Challenges, countermeasures, and future
Full
Gansen ZHAO1, 4, Wenfeng XU1, 4, Cheng QIAN1, 4, Zhihao HOU1, 4, Mengqin NING2, Yue GONG2, Guangyuan KONG1, 4, Xiangmin XU5, 6, Jiahong GUO2, *, Wenjun MA1, 3, *
Affiliations
  • 1School of Computer Science, South China Normal University, Guangzhou 510631, China
  • 2Institute of Logic and Cognitive Science, School of Philosophy, Beijing Normal University, Beijing 100875, China
  • 3Aberdeen Institute of Data Science and Artificial Intelligence, South China Normal University, Foshan 528225, China
  • 4Key Lab on Cloud Security and Assessment technology of Guangzhou, South China Normal University, Guangzhou 510631, China
  • 5Foshan University, Foshan 528225, China
  • 6Guangdong Artificial Intelligence and Digital Economy Laboratory (Guangzhou), Guangzhou 510335, China
Published: 2026-06-28 doi: 10.3981/j.issn.1000-7857.2025.05.00145
Outline
收藏切换

As a frontier technology in the current artificial intelligence era, generative artificial intelligence (GAI) is profoundly reshaping society across multiple domains. Concomitant with the proliferation of GAI applications are the emerging security challenges at both technical and social levels. To better understand the technical and social issues brought about by GAI, it is imperative to conduct a systematic and comprehensive investigation into existing security challenges, developed countermeasures, and future directions. This paper conducts a systematic survey spanning the entire lifecycle of GAI—namely, training security, inference security, and derived security—which correspond to model training, model inference, and model application, respectively. Formal models for GAI training and inference are developed to articulate security vulnerabilities, threat surfaces, and associated factors. An in−depth analysis covers typical attacks and corresponding countermeasures at both the training and inference stages. Derived security is also investigated, referring to security risks arising from GAI applications, including misinformation, social fairness concerns, individual privacy threats, and more, followed by a review of relevant countermeasures. Future directions in GAI security governance are discussed, emphasizing controllable and trustworthy generation mechanisms, security evaluation benchmarks, accountability mechanisms, and regulatory compliance frameworks. Overall, this article conducts a comprehensive investigation into the security risks and defense measures of GAI and constructs a security research framework that covers the entire lifecycle of GAI, as well as its research status at both technical and social levels.

generative artificial intelligence  /  training security  /  inference security  /  derivative security  /  social ethics
Gansen ZHAO, Wenfeng XU, Cheng QIAN, Zhihao HOU, Mengqin NING, Yue GONG, Guangyuan KONG, Xiangmin XU, Jiahong GUO, Wenjun MA. Security of generative artificial intelligence: Challenges, countermeasures, and future[J]. Science & Technology Review, 2026 , 44 (12) : 68 -86 . DOI: 10.3981/j.issn.1000-7857.2025.05.00145
Year 2026 volume 44 Issue 12
PDF
624
320
Cite this Article
BibTeX
Article Info
doi: 10.3981/j.issn.1000-7857.2025.05.00145
  • Receive Date:2025-05-27
  • Online Date:2026-07-17
  • Published:2026-06-28
Article Data
Affiliations
History
  • Received:2025-05-27
  • Revised:2026-04-22
Funding
Affiliations
    1School of Computer Science, South China Normal University, Guangzhou 510631, China
    2Institute of Logic and Cognitive Science, School of Philosophy, Beijing Normal University, Beijing 100875, China
    3Aberdeen Institute of Data Science and Artificial Intelligence, South China Normal University, Foshan 528225, China
    4Key Lab on Cloud Security and Assessment technology of Guangzhou, South China Normal University, Guangzhou 510631, China
    5Foshan University, Foshan 528225, China
    6Guangdong Artificial Intelligence and Digital Economy Laboratory (Guangzhou), Guangzhou 510335, China
References
Share
https://castjournals.cast.org.cn/joweb/kjdb/EN/10.3981/j.issn.1000-7857.2025.05.00145
Share to
QR

Scan QR to access full text

Cite this article
BibTeX
Citations
表12种不同金属材料的力学参数

Family
属数
Number of
genus
种数
Number of
species
占总种数比例
Percentage of
total species (%)

Genus
种数
Number of
species
占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏
  • BibTeX
  • EndNote
  • RefWorks
  • TxT