CRISC Certified in Risk and Information Systems ControlGovernanceHard
A startup company is rapidly developing a new AI-powered personal assistant. The product handles highly sensitive user data, including health information and financial transactions. The board of directors is concerned about the ethical implications of the AI's decision-making and data usage. Which of the following governance approaches should the company prioritize to instill trust and ensure responsible AI development?
- AForm a legal team to review potential liabilities.
- BImplement robust data encryption and access controls.
- CObtain industry-standard security certifications.
- DDevelop an 'ethics by design' framework for the AI system.
Show answer & explanationAnswer & explanation
Correct answer: D. Develop an 'ethics by design' framework for the AI system.
An 'ethics by design' framework proactively embeds ethical considerations into the AI system's development from its inception, addressing concerns about decision-making and data usage at a foundational level, which is critical for highly sensitive data and AI.
Why the other options are wrong
- A. A legal team primarily addresses liabilities reactively; it doesn't proactively build ethical decision-making into the AI system itself.
- B. Robust encryption and access controls are essential for security but do not address the ethical implications of the AI's decision-making or responsible data usage beyond security.
- C. Security certifications focus on technical security controls, not the broader ethical principles governing AI behavior and data handling.
Ethics by Design (EbD)
An approach that integrates ethical considerations and principles directly into the design, development, and deployment of systems, especially AI, from the earliest stages, ensuring that ethical values are embedded into the technology's architecture and functionality.
- Proactive, not reactive, ethical management.
- Aims to prevent ethical issues before they arise.
- Crucial for AI and systems handling sensitive data.
Memory trick: To build a trustworthy AI, bake ethics into its very code, don't just sprinkle it on top later.