A Survey of Agentic AI and Cybersecurity: Challenges, Opportunities and Use-case Prototypes

Обзор агентного искусственного интеллекта и кибербезопасности: проблемы, возможности и прототипы практических применений
Elisa Bertino, Maanak Gupta, Kshitiz Aryal, Sahaya Jestus Lazer
2026-01-08

agentic AIautonomous incident responsecybersecuritymemory poisoningthreat hunting
Agentic AI marks an important transition from single-step generative models to systems capable of reasoning, planning, acting, and adapting over long-lasting tasks. By integrating memory, tool use, and iterative decision cycles, these systems enable continuous, autonomous workflows in real-world environments. This survey examines the implications of agentic AI for cybersecurity. On the defensive side, agentic capabilities enable continuous monitoring, autonomous incident response, adaptive threat hunting, and fraud detection at scale. Conversely, the same properties amplify adversarial power by accelerating reconnaissance, exploitation, coordination, and social-engineering attacks. These dual-use dynamics expose fundamental gaps in existing governance, assurance, and accountability mechanisms, which were largely designed for non-autonomous and short-lived AI systems. To address these challenges, we survey emerging threat models, security frameworks, and evaluation pipelines tailored to agentic systems, and analyze systemic risks including agent collusion, cascading failures, oversight evasion, and memory poisoning. Finally, we present three representative use-case implementations that illustrate how agentic AI behaves in practical cybersecurity workflows, and how design choices shape reliability, safety, and operational effectiveness.
1
Agentic AI extends generative models with memory, tool use, reasoning, planning, action, and adaptation across long-running tasks.
2
Agentic capabilities also strengthen adversaries by accelerating reconnaissance, exploitation, coordinated operations, and social-engineering attacks.
3
Agentic cybersecurity systems face systemic risks including agent collusion, cascading failures, oversight evasion, and memory poisoning; three use-case implementations illustrate how design choices affect reliability, safety, and operational effectiveness.
4
Existing governance, assurance, and accountability mechanisms have fundamental gaps because they were designed for non-autonomous, short-lived AI systems.
5
In cybersecurity defense, agentic systems support continuous monitoring, autonomous incident response, adaptive threat hunting, and large-scale fraud detection.

agentic AI systems in cybersecurity workflows

their defensive and adversarial capabilities, systemic risks, and governance, assurance, reliability, safety, and operational effectiveness

Publication Details
Publication Date
2026-01-08
Journal
Publisher
ISSN
Cited by
56
Access Type
Author Information
Authors
Elisa Bertino
Maanak Gupta
Kshitiz Aryal
Sahaya Jestus Lazer
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%