Review Article
The Dual Use of Generative Artificial Intelligence in Cybersecurity: A Survey from Deepfake Attacks to Intelligent Defense
Dalin Xiang*
,
Xinkai Fu
Issue:
Volume 1, Issue 3, September 2026
Pages:
112-116
Received:
11 September 2026
Accepted:
20 September 2026
Published:
30 September 2026
Abstract: The rapid iteration of generative Artificial Intelligence (AI) is reshaping the cybersecurity offense-defense landscape. This paper systematically reviews research progress in this field from 2021 to 2026, focusing on the dual use of generative AI in cybersecurity. On the one hand, malicious actors exploit generative AI to create deepfake content, automate phishing emails, and conduct social engineering attacks, significantly improving the personalization and scalability of attacks. On the other hand, the same technological system demonstrates significant effectiveness in defense scenarios such as threat detection, vulnerability remediation, data augmentation, and security knowledge management. Existing reviews often treat attack and defense separately, lacking an integrated analytical framework. This paper analyzes the field from four dimensions: attack surface expansion, defense technology evolution, governance frameworks, and evaluation benchmarks. We first review generative AI-driven cyberattacks, covering deepfake attacks, automated phishing, automated malware development, and intelligent social engineering. We then examine defense applications, including intelligent threat detection, data augmentation and class imbalance handling, vulnerability detection and automated remediation, security knowledge management, and security education. Next, we identify key shortcomings in current research regarding adversarial robustness, privacy protection, interpretability, and deployment constraints. We further outline five future research directions: adaptive defense and adversarial robustness, federated learning and privacy protection, interpretability and human-machine collaboration, interdisciplinary governance frameworks, and multimodal security evaluation benchmarks. This review aims to provide a structured reference for building adaptive defense systems and interdisciplinary governance paths in the era of generative AI.
Abstract: The rapid iteration of generative Artificial Intelligence (AI) is reshaping the cybersecurity offense-defense landscape. This paper systematically reviews research progress in this field from 2021 to 2026, focusing on the dual use of generative AI in cybersecurity. On the one hand, malicious actors exploit generative AI to create deepfake content, ...
Show More
Review Article
Teaching Students to Verify Generative AI Using the TRACE-AI Framework
Issue:
Volume 1, Issue 3, September 2026
Pages:
117-129
Received:
3 August 2026
Accepted:
17 August 2026
Published:
9 November 2026
DOI:
10.11648/j.sdai.20260103.12
Downloads:
Views:
Abstract: Generative artificial intelligence (GenAI) can produce polished mathematical solutions, executable code, persuasive media, and professional recommendations, yet fluency does not establish correctness, provenance, security, or contextual fitness. Existing AI-literacy frameworks appropriately describe broad competencies, while emerging studies in mathematics, programming, and health professions education show the value of domain-specific validation. What remains underdeveloped is a portable instructional process that makes verification, uncertainty, and reliance observable across disciplines. This conceptual article introduces TRACE-AI: Translate the task and stakes; Represent the problem, assumptions, and constraints; Audit the AI's reasoning and output; Corroborate pivotal claims through independent evidence, tests, or tools; Estimate uncertainty and warranted reliance; Act proportionately by accepting, revising, rejecting, or escalating; and Internalize learning through reflection and unaided transfer. The framework was developed through a theory-informed critical synthesis of AI literacy, mathematical problem solving, computational thinking and debugging, epistemic cognition, self-regulated learning, and trust calibration. Each stage produces an assessable artefact, and a risk gate increases verification and human oversight as uncertainty, irreversibility, or potential harm rises. Worked applications in mathematics, programming/cybersecurity, and health-professional reasoning demonstrate invariant functions with discipline-specific standards of proof. An assessment architecture distinguishes final-answer quality from process quality, appropriate reliance, confidence calibration, verification quality, and delayed unaided transfer. Six propositions define a research programme for comparative and randomized evaluation. TRACE-AI is not presented as an effectiveness-tested intervention; it is a transdisciplinary curriculum and research architecture requiring contextual adaptation and empirical validation. Its contribution is to move GenAI education beyond prompt production toward disciplined human judgment about when AI output deserves trust.
Abstract: Generative artificial intelligence (GenAI) can produce polished mathematical solutions, executable code, persuasive media, and professional recommendations, yet fluency does not establish correctness, provenance, security, or contextual fitness. Existing AI-literacy frameworks appropriately describe broad competencies, while emerging studies in mat...
Show More
