Safety and Security of AI Systems and Software
Albert J. Marcella · 2025
This chapter examines the safety and security considerations for artificial intelligence (AI) systems, emphasizing the need for robust internal controls, regulatory compliance, and proactive risk management. It explores how organizations can safeguard AI systems against potential cyber threats, adversarial attacks, and failures, ensuring their reliability and accountability. The chapter underscores the significance of maintaining secure data pipelines, enforcing ethical usage, and conducting comprehensive security audits. It incorporates guidelines from frameworks like the AI Bill of Rights to establish standards for safe and effective AI deployment. Audit questions are provided to assess the system’s security protocols, data privacy compliance, and incident response mechanisms, ensuring AI systems operate within defined safety limits and meet organizational expectations.