Security, Compliance, and Governance for AI Solutions

Question 1

What’s the difference between Security, Governance, and Compliance?

Answer 1

Security: Ensure that confidentiality, integrity, and availability are maintained for organizational data and information assets and infrastructure. This function is often called information security or cybersecurity in an organization.

Governance: Ensure that an organization can add value and manage risk in the operation of business.

Compliance: Ensure normative adherence to requirements across the functions of an organization.

Question 2

What does Governance typically involve?

Answer 2

1. Board-level or senior management ownership
2. Clearly defined roles and responsibilities
3. Policies and procedures.
4. Training and awareness

Question 3

Why do you need governance in an organization?

Answer 3

1. Enables you to manage, optimize, and scale AI initiatives while minimizing risks and downsides.
2. Increases transparency and trustworthiness
3. Enables regulatory compliance.

Question 4

What makes AI compliance challenging?

Answer 4

1. Complexity and opaque - LLMs are hard to audit as they are neural network based and the decision making process (explainability) is difficult to understand
2. Dynamism - they change over time as they learn
3. Emergent capabilities - new unforeseen capabilities
4. Unique risks not seen in traditional systems (e.g. data and algo bias)
5. Algorithm accountability

Question 5

What does data governance in AI involve?

Answer 5

Overall, it is applicable to the entire data lifecycle from its collection and storage, to usage and security. For AI, it involves:
a) Data quality and integrity - complete and accurate, validation, cleansing, addressing anomalies and inconsistencies, maintaining data lineage and provenance

b) Data protection and privacy - develop and enforce policies to protect sensitive and personal info, access control, encryption, incident response

c) Data lifecycle management - classify and catalog data, data retention and disposition, data backup and recovery

d) Responsible AI - bias, fairness, transparency, and accountability,

e) Governance structures and roles

f) Data sharing and collaboration - develop data sharing agreements and protocols

Question 6

What is data logging in the context of AI?

Answer 6

1. Tracking inputs and outputs
2. Model performance metrics
3. Capturing System events.