Abstract: The rapid adoption of artificial intelligence (AI) in cybersecurity has created a double-edged sword: while AI enhances threat detection and response capabilities, it also creates new vulnerabilities through adversarial attacks and raises serious ethical concerns regarding fairness, accountability, and transparency. Current approaches often view technical security and ethical governance as separate challenges, and produce isolated solutions that are inadequate against emerging AI-driven threats. This paper proposes a new integrated ethical-security framework (IESF) that bridges this gap by incorporating ethical governance directly into AI cybersecurity operations. The framework offers a multi-layered architecture that combines technical threat analysis with real-time ethical impact assessments, dynamic human-AI collaboration protocols, and comprehensive accountability mechanisms. The results show 95.7% detection accuracy (11.4% improvement), 74.4% reduction in false positives, and 71.1% lower success rate for adversarial attacks. Ethically, this framework achieved 98% bias detection, 96.5% privacy protection, and 99.8% audit trail completeness. Operationally, the system reduced human analyst workload by 78.5% while improving response times by 79%, demonstrating that ethical integration enhances rather than hinders security effectiveness. This research provides both a practical framework for responsible AI cybersecurity and empirical evidence that integrating ethical considerations with technical security produces synergistic benefits. The findings show that the IESF provides a viable path for organizations to harness the defensive potential of AI while ensuring alignment with human values, providing a significant advance toward reliable and effective cybersecurity in the age of AI.

Key words: AI cybersecurity, ethical AI, adversarial machine learning, AI governance, human-AI collaboration