Artificial Intelligence (AI) Trust Framework and Maturity Model: Applying an Entropy Lens to Improve Security,...

Area of Science:

• Computer Science
• Information Theory
• Artificial Intelligence Ethics

Background:

• Rapid AI advancements necessitate robust ethical, moral, and legal safeguards.
• Current metrics for AI security and privacy are insufficient for ethical AI management.
• Lack of transparency in
• black box
• AI systems hinders trust and ethical oversight.

Purpose of the Study:

• Propose an AI Trust Framework and Maturity Model to enhance trust in AI system design and management.
• Develop metrics for assessing AI security, privacy, and ethical performance.
• Improve transparency and trust in AI, especially in complex human-machine systems.

Main Methods:

• Utilized an
• entropy lens
• rooted in information theory to analyze AI systems.
• Developed a maturity model for assessing AI trust.
• Applied the framework to two use cases for validation.

Main Results:

• The proposed framework enhances transparency and trust in AI systems.
• The
• entropy lens
• provides a quantifiable measure for AI trust.
• Demonstrated improved trust and performance in human-machine teams through the framework.

Conclusions:

• The AI Trust Framework and Maturity Model offers a novel approach to ethically manage AI.
• Quantifying AI trust through entropy can optimize performance in autonomous systems and teams.
• The framework is validated through practical use cases, showing its utility in AI design and management.