Open Access

Adaptive Identity Fabric for Agentic AI Systems: Architecture, Trust Management, and Security Evaluation

https://doi.org/10.64917/feaiml/Volume03Issue09-01

PDF

Kenji Mori ¹ Aiko Fujimoto ¹

⁴ Machine Learning Research Specialist, Japan

⁴ AI Systems Research Analyst, Japan

Abstract

The emergence of agentic artificial intelligence (AI) systems introduces a security problem that differs fundamentally from conventional application authentication. Autonomous agents can initiate actions, invoke tools, exchange information, delegate subtasks, and operate across heterogeneous infrastructure without continuous human intervention. Consequently, identity must become a continuously evaluated security property rather than a static authentication attribute. This paper proposes an Adaptive Identity Fabric (AIF) for agentic AI systems, integrating workload identity, contextual authentication, trust evaluation, authorization, identity lifecycle management, and security telemetry into a unified architecture. The methodology is conceptually grounded in adaptive representation, feature selection, multiscale processing, and learning-oriented system design reflected in the supplied literature, while the principal identity-fabric direction is positioned using Pappu, Bhushan, and Jaiswal (2026). The proposed framework separates identity establishment from trust assessment and authorization, enabling agent identities to adapt according to workload context, behavioral evidence, resource sensitivity, and trust changes. A conceptual security evaluation examines resilience against identity spoofing, privilege escalation, compromised agents, unauthorized delegation, and trust drift. Findings indicate that an adaptive identity fabric can provide stronger security than static identity mechanisms by continuously correlating identity, context, behavior, and policy. However, the approach introduces computational overhead, policy complexity, trust-calibration challenges, and dependency on reliable telemetry. The study contributes an architectural model and evaluation framework for identity-centric security in autonomous AI ecosystems.

How to Cite

Kenji Mori, & Aiko Fujimoto. (2026). Adaptive Identity Fabric for Agentic AI Systems: Architecture, Trust Management, and Security Evaluation. Frontiers in Emerging Artificial Intelligence and Machine Learning, 3(09), 01–06. https://doi.org/10.64917/feaiml/Volume03Issue09-01

⬇ Endnote/Zotero/Mendeley (RIS) ⬇ BibTeX

References

J. Bai, B. Gong, Y. Zhao, F. Lei, C. Yan, and Y. Gao, “Multi-scale representation learning on hypergraph for 3D shape retrieval and recognition,” IEEE Trans. Image Process., vol. 30, pp. 5327–5338, May 2021.

Z. Cao, K. Zhang, and J. Wu, “FPB: Improving multi-scale feature representation inside convolutional layer via feature pyramid block,” in Proc. IEEE Int. Conf. Image Process., 2020, pp. 1666–1670.

S. Chen, Y. Cao, Y. Kang, P. Li, and B. Sun, “Deep feature representation based imitation learning for autonomous helicopter aerobatics,” IEEE Trans. Artif. Intell., vol. 2, no. 5, pp. 437–446, Oct. 2021.

M. N. Do and M. Vetterli, “The finite ridgelet transform for image representation,” IEEE Trans. Image Process., vol. 12, no. 1, pp. 16–28, Jan. 2003.

Y. Duan, F. Liu, L. Jiao, P. Zhao, and L. Zhang, “SAR image segmentation based on convolutional-wavelet neural network and Markov random field,” Pattern Recognit., vol. 64, pp. 255–267, 2017.

L. Jiao, B. Hou, W. Shuang, and L. Fang, Image Multiscale Geometric Analysis: Theory and Applications. Xi’an, China: Xian Electron. Sci. Technol. Univ. Press, 2008.

L. Jiao, J. Pan, and Y. Fang, “Multiwavelet neural network and its approximation properties,” IEEE Trans. Neural Netw., vol. 12, no. 5, pp. 1060–1066, Sep. 2001.

L. Jiao, Neural Network Computing. Xi’an, China: Xidian Univ. Press, 1993.

L. Jiao, Neural Network System Theory. Xi’an, China: Xidian Univ. Press, 1990.

L. Jiao, S. Yang, F. Liu, S. Wang, and Z. Feng, “Seventy years of neural networks: Review and prospect,” Chin. J. Comput., vol. 39, no. 8, pp. 1697–1716, 2016.

Y. W. Liyanage, D.-S. Zois, and C. Chelmis, “Dynamic instance-wise joint feature selection and classification,” IEEE Trans. Artif. Intell., vol. 2, no. 2, pp. 169–184, Apr. 2021.

D. Luengo, D. Meltzer, and T. Trigano, “Sparse ECG representation with a multi-scale dictionary derived from real-world signals,” in Proc. 41st Int. Conf. Telecommun. Signal Process., 2018, pp. 1–5.

J. Shi, Y. Zhao, W. Xiang, V. Monga, X. Liu, and R. Tao, “Deep scattering network with fractional wavelet transform,” IEEE Trans. Signal Process., vol. 69, pp. 4740–4757, Jul. 2021.

J. Yu, M. Tan, H. Zhang, D. Tao, and Y. Rui, “Hierarchical deep click feature prediction for fine-grained image recognition,” IEEE Trans. Pattern Anal. Mach. Intell., to be published, doi: 10.1109/TPAMI.2019.2932058.

L. Zhang, W. Zhou, and L. Jiao, “Wavelet support vector machine,” IEEE Trans. Syst., Man, Cybern., Part B. (Cybern.), vol. 34, no. 1, pp. 34–39, Feb. 2004.

D. Zhou, R. Duan, L. Zhao, and X. Chai, “Single image super-resolution reconstruction based on multi-scale feature mapping adversarial network,” Signal Process., vol. 166, 2020, Art. no. 107251.

K. Pappu, B. Bhushan and N. Jaiswal, "Future-Proofing Identity Security for Agentic AI Systems: Design, Implementation, and Evaluation of an Identity Fabric," 2026 International Conference on Artificial Intelligence, Systems, and Emerging Technologies (ICAISET), Cairo, Egypt, 2026, pp. 1-7, doi: 10.1109/ICAISET66439.2026.11541783.