What does 'Security by Design' mean in AI cybersecurity programs?

• A. Integrating security measures from the outset when adapting or creating new programs.
• B. Adding security features after deployment only when problems are reported.
• C. Focusing exclusively on user experience over security.
• D. Outsourcing all security to third parties.

Answer: (A)

Security by design means weaving security into every stage of creating and adapting AI programs, not tacking it on after the fact. In practice this means starting with threat modeling and security requirements, choosing architectures that limit risk, and implementing controls from the ground up—such as secure coding, input validation, encryption, strong authentication, access control, and secure data handling. For AI specifically, it includes protecting data used for training and inference, ensuring model integrity and provenance, guarding against data poisoning and adversarial manipulation, and planning for secure updates and incident response. Ongoing monitoring, testing, and governance are part of the lifecycle so vulnerabilities are caught early and accountability is maintained. This approach reduces the attack surface, lowers remediation costs, and helps meet privacy and regulatory expectations from the outset. Reactive security after deployment, prioritizing user experience over security, or outsourcing security entirely miss this integrated, proactive mindset.