Work

Examining the Role of Artificial Intelligence in Strengthening Cyber Threat Intelligence Across U.S. Public Sector IT Infrastructure

Year: 2025

Type: article

Abstract: Cyber threats targeting the U.S. public sector have escalated beyond the capacity of traditional security measures. Government agencies now face heightened vulnerability as they process vast amounts of sensitive data through increasingly complex technological infrastructures. This study investigates how artificial intelligence can enhance cyber threat intelligence throughout U.S. federal, state, and local government networks. A systematic literature review approach was employed, and peer-reviewed journal publications were searched across scientific databases, including IEEE Xplore, ScienceDirect, Scopus, and Web of Science, for government-enabled cybersecurity AI applications. The findings demonstrate that machine learning and deep learning capabilities dramatically increase the accuracy of threat detection, with organizations reporting a 75% reduction in the number of breaches because of implementing AI compared to those that do not rely on this technology. The quantitative metrics showed that AI-augmented models achieve detection rates with average accuracy scores of 1.000, successfully surpassing traditional signature-based techniques in uncovering advanced persistent threats and zero-day attacks. Notwithstanding, challenges such as budget considerations for 65.75% of institutions, IT workforce shortages for 55.25%, and the insecurities that come with integrating legacy systems into digital ones exist. This study establishes that the effective deployment of AI must be predicated on transparent training algorithms, greater inter-agency cooperation, adequate funding to modernize technical capabilities, train personnel, and compliance with regulatory frameworks, including FISMA, NIST standards, and adherence to the zero-trust architectural model, which are necessary prerequisites for robustly defending critical national infrastructure.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.18398805

Authors: Mahmoud, Mariatu, Aryee, Barbara Aryeley, Agyemang, Kwadwo Adu

Institution: East Tennessee State University

Language: English

Topic: Network Security and Intrusion Detection

Subfield: Computer Networks and Communications

Field: Computer Science

Domain: Physical Sciences

SDG: Industry, innovation and infrastructure

Open Access status: green

OpenAlex ID: W7125975749

FWCI:

Cites:

Cited by:

Related to: