Systematic review finds adaptive AI-cybersecurity strategies and workforce gaps define national policy options

A systematic review of 53 studies published from 2016 through February 2025 found broad agreement that artificial intelligence can support cyberthreat detection, automated response and vulnerability management, while also creating security and ethical risks. The review assessed both AI for cybersecurity and cybersecurity for AI, alongside national strategies, economic implications and policy recommendations across the United States, European Union, China, UAE and Africa, including Ethiopia.

The researchers followed a structured literature-review design informed by the PRISMA 2020 reporting guidelines. Their search covered IEEE Xplore, ScienceDirect, the ACM Digital Library, Springer Nature and Google Scholar, using terms for artificial intelligence, machine learning and deep learning alongside cybersecurity, intrusion detection, threat detection, policy, strategy, governance and initiatives. The search returned 2,758 records; an open-access filter left 192, duplicate removal reduced the pool to 93, and title, abstract and full-text screening resulted in 53 included articles, books, theses and scientific reports.

Across the included research, machine learning, deep learning and natural-language-processing methods were associated with applications in network-intrusion detection, traffic analysis, malware identification, phishing detection, fraud analysis, incident response and vulnerability management. The review described adaptive models that learn from data as useful for identifying previously unseen attacks, while automation could reduce repetitive analytical work and shorten response times. These were findings from a heterogeneous literature review rather than results from a new controlled deployment or an efficacy meta-analysis of one security system.

The review also emphasized the reverse relationship: AI systems require cybersecurity protections for their training data, model-development pipelines, storage, networks and computational infrastructure. Proposed controls included access controls, encryption, data validation, integrity checks and secure data management. It categorized adversarial attacks according to attackers’ access to model information, including white-box, black-box and gray-box attacks. The literature also documented risks from data poisoning, AI-enabled phishing and malware, synthetic media, false positives, biased or incomplete datasets, privacy violations and the opacity of some models.

National strategies differed substantially. The review reported that the United States, China and the European Union had developed prominent national or regional AI initiatives, while the EU placed greater emphasis on individual rights and regulatory compliance and China placed greater emphasis on national security and public safety. The UAE had an AI Strategy 2031, while more than 40 countries and regions had reportedly elevated AI development to a national-strategy level since 2016. The review also noted that 80 nations had developed national cybersecurity strategies tailored to their environments and threats, although differences in policy priorities complicate international coordination.

Economic effects were described as potentially mixed. Automation and more efficient threat detection may produce cost savings, but initial investment and the need for specialized staff can be substantial barriers for smaller organizations and resource-constrained countries. The review cited market research projecting growth of just over 7% annually for artificial intelligence in cybersecurity, but it did not conduct an independent forecast or quantify net economic benefits. In Africa, identified constraints included limited technical expertise, restricted budgets, shortages of structured and high-quality data, inadequate infrastructure, incomplete laws and weak implementation capacity.

Ethiopia was examined as an emerging-economy case. The review noted the launch of the Digital Ethiopia 2030 national digital-transformation strategy in June 2020 and described Ethiopia’s AI sector as small, with companies and institutions including EthioCloud, GebeyaNet, Botter, Ethio Robo Robotics and iCog-Labs. It reported that Ethiopia’s strategy did not include a specific AI-cybersecurity implementation plan. The review also cited concerns involving cyberattacks against critical infrastructure, data protection, national sovereignty, political effects and the need to build institutional capacity.

The strongest recurring findings concerned implementation. In 47 of the 53 studies, authors favored adaptive, forward-looking defenses over static security controls, while 34 of the 53 studies identified shortages of professionals with combined AI and cybersecurity expertise as the most frequently cited barrier. The review’s recommendations included privacy-by-design, bias audits, explainable systems, human oversight, workforce education, public-private partnerships, cross-border policy coordination and proactive plans for emerging threats. It also proposed future work on quantum-resistant cybersecurity, lightweight systems for resource-constrained settings, standardized benchmarks and underrepresented regions.

The findings should be interpreted within several limitations. The search was limited to English-language publications, excluded conference proceedings and ended in February 2025. The included studies were qualitatively heterogeneous, and the researchers did not perform a statistical meta-analysis. The authors also described their work as following systematic-review principles and PRISMA 2020 for transparent reporting, while noting that a full systematic literature review would require resources beyond the study’s scope. The review therefore offers a broad synthesis of technical, policy and economic evidence, not a comparative test showing that any particular AI cybersecurity system outperforms another.

Subject of Research: Systematic review of bidirectional artificial intelligence and cybersecurity integration, including technical applications, national strategies, economic implications and policy limits.

Article Title: A systematic review of artificial intelligence and cybersecurity integration covering national strategies, economic implications, and policy recommendations

Article References: Musa, H. A. (2026). A systematic review of artificial intelligence and cybersecurity integration covering national strategies, economic implications, and policy recommendations. Discover Artificial Intelligence, 6(1), Article 1455. https://doi.org/10.1007/s44163-026-02309-6

Image Credits: AI Generated

DOI: 10.1007/s44163-026-02309-6

Keywords: artificial intelligence, cybersecurity, systematic review, national strategy, cyberthreat detection, AI security, information governance, workforce skills, Africa, Ethiopia, systematic, review