TY - GEN
T1 - An Access Control Method Against Unauthorized and Noncompliant Behaviors Leveraging Large Language Models
AU - Farhadighalati, Nastaran
AU - Kalateh, Sepideh
AU - Estrada-Jimenez, Luis A.
AU - Nikghadam-Hojjati, Sanaz
AU - Barata, José
N1 - Funding Information:
This research was partly supported by EU Funded project CAIOC with Grant Agreement ID: 101190339, the Portuguese FCT program, Centre of Technology and Systems (CTS) CTS/00066, and Nova University Lisbon’s PhD program of Department of Electronics and Computer Engineering.
Publisher Copyright:
© IFIP International Federation for Information Processing 2025.
PY - 2025/6/26
Y1 - 2025/6/26
N2 - In today’s digital world, protecting sensitive data is crucial for organizations. Traditional access control systems, which rely on fixed roles and attributes, often fail to adapt to evolving security threats. Behavior-Based Access Control (BBAC) addresses the shortcomings of conventional methods by dynamically adjusting user permissions in real time based on observed actions. To enhance BBAC, Large Language Models (LLMs) are incorporated, leveraging advanced language processing to interpret access policies, analyzing behavior patterns, and making context-aware decisions. Integrating Large Language Models (LLMs) enables the system to detect suspicious activities, such as unauthorized actions or policy violations, with greater accuracy. Consequently, BBAC systems achieve flexible and intelligent access control, balancing security with workflow efficiency. Early experiments using synthetic data confirm the model’s effectiveness. The experimental results demonstrate that the model is efficient, accurate, and shows potential in detecting abnormal behaviors among authorized users.
AB - In today’s digital world, protecting sensitive data is crucial for organizations. Traditional access control systems, which rely on fixed roles and attributes, often fail to adapt to evolving security threats. Behavior-Based Access Control (BBAC) addresses the shortcomings of conventional methods by dynamically adjusting user permissions in real time based on observed actions. To enhance BBAC, Large Language Models (LLMs) are incorporated, leveraging advanced language processing to interpret access policies, analyzing behavior patterns, and making context-aware decisions. Integrating Large Language Models (LLMs) enables the system to detect suspicious activities, such as unauthorized actions or policy violations, with greater accuracy. Consequently, BBAC systems achieve flexible and intelligent access control, balancing security with workflow efficiency. Early experiments using synthetic data confirm the model’s effectiveness. The experimental results demonstrate that the model is efficient, accurate, and shows potential in detecting abnormal behaviors among authorized users.
KW - Behavior-Based Access Control Model
KW - Data
KW - Large Language Model
KW - Privacy Protection
UR - https://www.scopus.com/pages/publications/105010034438
U2 - 10.1007/978-3-031-97051-1_3
DO - 10.1007/978-3-031-97051-1_3
M3 - Conference contribution
AN - SCOPUS:105010034438
SN - 9783031970504
T3 - IFIP Advances in Information and Communication Technology
SP - 39
EP - 50
BT - Technological Innovation for AI-Powered Cyber-Physical Systems
A2 - Camarinha-Matos, Luis M.
A2 - Ferrada, Filipa
PB - Springer, Cham
CY - Switzerland
T2 - 16th IFIP WG 5.5 / SOCOLNET Advanced Doctoral Conference on Computing, Electrical, and Industrial Systems, DoCEIS 2025
Y2 - 2 July 2025 through 4 July 2025
ER -