TY - GEN
T1 - A Survey on Trustworthy Systems in ChatGPT-Like Large-Scale Generative AI Models
T2 - 7th International Conference on Blockchain, Artificial Intelligence, and Trustworthy Systems, BlockSys 2025
AU - Li, Dun
AU - Li, Hongzhi
AU - Li, Jiatao
AU - Crespi, Noel
AU - Minerva, Roberto
AU - Li, Kuan Ching
AU - Hu, Lingxiang
AU - Shao, Wenhao
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2026.
PY - 2026/1/1
Y1 - 2026/1/1
N2 - Large-scale generative AI models such as ChatGPT perform well in text generation, automatic dialogue, and multimodal tasks. However, the widespread deployment of these systems presents challenges related to their credibility and trustworthiness, encompassing issues such as security vulnerabilities, interpretability limitations, reliability concerns, and privacy risks. This study systematically examines the latest advancements in dealing with these challenges through a comprehensive, interdisciplinary approach. By integrating insights from technology, ethics, and policy domains, the research underscores the multifaceted implications of deploying large-scale generative AI systems. In addition, this paper provides practical insight into a structured framework centered around the core pillars of trusted AI systems, guiding the design, development, and deployment of generative AI models that meet societal expectations, helping to improve understanding of the credibility of generative AI, and promoting its responsible integration in various applications.
AB - Large-scale generative AI models such as ChatGPT perform well in text generation, automatic dialogue, and multimodal tasks. However, the widespread deployment of these systems presents challenges related to their credibility and trustworthiness, encompassing issues such as security vulnerabilities, interpretability limitations, reliability concerns, and privacy risks. This study systematically examines the latest advancements in dealing with these challenges through a comprehensive, interdisciplinary approach. By integrating insights from technology, ethics, and policy domains, the research underscores the multifaceted implications of deploying large-scale generative AI systems. In addition, this paper provides practical insight into a structured framework centered around the core pillars of trusted AI systems, guiding the design, development, and deployment of generative AI models that meet societal expectations, helping to improve understanding of the credibility of generative AI, and promoting its responsible integration in various applications.
KW - ChatGPT
KW - Generative AI
KW - Interpretability
KW - Reliability
KW - Security
KW - Trustworthy AI
UR - https://www.scopus.com/pages/publications/105028290079
U2 - 10.1007/978-981-95-3480-7_18
DO - 10.1007/978-981-95-3480-7_18
M3 - Conference contribution
AN - SCOPUS:105028290079
SN - 9789819534791
T3 - Communications in Computer and Information Science
SP - 249
EP - 260
BT - Blockchain and Trustworthy Systems - 7th International Conference on Blockchain, Artificial Intelligence, and Trustworthy Systems, BlockSys 2025, Revised Selected Papers
A2 - Chen, Jianguo
A2 - Luo, Xiaonan
A2 - Yu, Yuanlong
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 30 May 2025 through 31 May 2025
ER -