IT Examiner School - Oct 2025

A.I. Risk Management Considerations Governance Data Security Transparency Reliability

Fairness and Bias

• No formal policy with defined roles in AI decision-making and risks oversight. •Weak governance structures may struggle to identify, monitor, and mitigate risks effectively.

• Sensitive data and weakly secured AI models can create vulnerabilities, leading to data exposure or manipulation through data poisoning and adversarial attacks, which

• Feeding

• Complex AI

• Due to changing data patterns, AI systems may become less accurate over time, leading to poor decision making.

algorithms with incomplete or incorrect data is the primary cause of unintended bias in AI outputs.

algorithms often lack explainability, making it difficult to understand or justify decisions.

can compromise system integrity.

A.I. Risk Management Framework

• Board • Executives

Creates

Reviews

• Risk Managers • AI Developers • Users

Management Process

AI Model (Black Box)

Results

Inputs

• Risk Tolerance • Risk Assessment • Policies • Procedures • Reporting

Controls

Reports

Analysis