BEYOND AUTOMATION: GENERATIVE AI’S INFLUENCE ON HUMAN DECISION-MAKING, TRUST, AND ACCOUNTABILITY IN INTELLIGENT CYBERSECURITY SYSTEMS

Abstract

Generative artificial intelligence (GenAI) is transforming cybersecurity by enabling security professionals to interpret complex threat information, summarize security alerts, generate incident-response strategies, prioritize vulnerabilities, and support organizational risk management. Unlike conventional automation, which primarily executes predefined rules and procedures, GenAI generates contextual explanations and recommendations that can influence how humans interpret evidence and make consequential decisions. These capabilities offer opportunities to improve analytical efficiency, reduce cognitive workload, and strengthen cyber defense. However, they also introduce challenges involving automation bias, inappropriate trust, hallucinated information, insufficient transparency, and unclear accountability. This conceptual research article examines the relationship between GenAI, human decision-making, trust calibration, and accountability in intelligent cybersecurity systems. Using a theory-driven narrative synthesis, the article develops an integrated framework connecting technological capabilities, human interpretation, verification practices, decision quality, and organizational governance. The framework identifies system reliability, professional expertise, workload, evidence accessibility, and governance maturity as factors influencing human–AI collaboration. Eight analytical tables summarize technological distinctions, conceptual constructs, research questions, cybersecurity applications, governance mechanisms, research propositions, and implementation priorities. Seven propositions are advanced for future empirical testing. The article argues that GenAI's value depends not simply on its ability to automate tasks but on whether its recommendations remain verifiable, its limitations are understood, and human professionals retain meaningful authority over consequential actions. A risk-sensitive governance framework is proposed, incorporating evidence-linked explanations, least-privilege permissions, human approval thresholds, auditable decision records, continuous monitoring, and organizational learning. The article concludes that responsible cybersecurity requires a transition from automation-centered deployment toward accountable human–AI collaboration, in which generative systems strengthen human analytical capacity without weakening professional judgment or institutional responsibility.