CoChat
Sign Up
Cited by 9Network Security and Intrusion DetectionNetwork Packet Processing and OptimizationAdvanced Malware Detection Techniques

Agentic AI for Autonomous Cyber Threat Hunting and Adaptive Defense in Dynamic Security Environments

Sheth, Adit, Patel, Advait, Upadhyay, Charit, Ragothaman, Hariharan, Patil, Balkrishna, Udayakumar, Saai Krishnan·2025·2025 IEEE International Conference on Electro Information Technology (eIT)
DOI 10.1109/eit64391.2025.11103697
Open this paper in CoChat — free account

Abstract

Evolution of cyber threats have led to limitations in the efficiency of traditional rule-based cybersecurity approaches. These traditional interventions have remained to have a resource intensive and reaction approach, prompting the creation of new steps that can help in handling cyber threats through AI. This study develops an AI driven system for autonomous cyber threat hunting, through leveraging Machine Learning (ML), Deep Reinforcement Learning (RDL) and using AI analytics to ensure a proactive detection, resolution and management of threats. The study uses datasets such as DARPA and CIDS ensuring that they can work on an AI intervention. The use of ML models such as CNN and RNN ensure analysis of the network traffic, behavioral indicators and system logs. Using DRL enables the system to have an autonomous adaptation to threats such as zero day exploits, reducing the instances of false positives and false negatives. Results from this study indicates that using AI driven threat detection leads to an accuracy rate of 98.2%, F1 score of 97.1% while reducing both MTTD and MTTR by an estimated 99%. The findings depict the ability of the system to have an operational framework demanding minimal human intervention. Thus, the cybersecurity operations can be advanced to have an increasingly transformative engagement, working on proactive threat detection and reducing operational costs. Considerably, future research should focus on enhancing Explainability in AI, addressing ethical issues and countering growing adversarial attacks on the AI system.

Integrity signals

Registry

Crossref · 10.1109/eit64391.2025.11103697

Cites

14 of 14 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Claims the paper makes

We couldn’t retrieve this paper’s full text, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Agentic AI systems wrap LLM reasoning in perception, planning, memory, tool-use, and multi-agent communication loops, enabling autonomous execution of multi-step security tasks.

    “The second is the development of agentic AI: software systems that wrap LLM reasoning in perception, planning, memory, tool-use, and multi-agent communication loops, enabling them to autonomously execute multi-step security tasks that previously required human analysts”

    From Agentic AI and Large Language Models for Autonomous IoT Cybersecurity: A Systematic Survey, Taxonomy, and Research Roadmap (2026) · cites this paper as [24]

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Sheth, A., Patel, A., Upadhyay, C., Ragothaman, H., Patil, B., & Udayakumar, S. K.
Affiliations
  • Microsoft Corporation
  • Advanced Micro Devices (United States) (US)
  • Salesforce (United States) (US)
ORCID
Venue
2025 IEEE International Conference on Electro Information Technology (eIT)
Publisher
IEEE
Published
2025-05-29
Pages
316-321
ISSN
2154-0357 (ISSN-L)
DOI
10.1109/eit64391.2025.11103697
OpenAlex
W4413179320
Cited by
7
References
14
Citations by year
2026: 5 · 2025: 2
FWCI
8.72
Type
conference paper
Language
en
Open access
closed
Licence
Locations
IEEE International Conference on Electro Information Technology (publishedVersion)
Topics
  • Physical Sciences › Computer Science › Computer Networks and Communications › Network Security and Intrusion Detection
  • Physical Sciences › Computer Science › Hardware and Architecture › Network Packet Processing and Optimization
  • Physical Sciences › Computer Science › Signal Processing › Advanced Malware Detection Techniques
Keywords
agentic AI, deep reinforcement learning, network traffic analysis, zero-day attacks, false positive reduction, mean time to detect, proactive threat detection, explainability

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review