AI Assurance Logo
AI Assurance aiassurance.co.za
Get In Touch

King V and AI:
What boards must ask before
approving the next GenAI pilot

New King V guidance places heightened responsibility on directors for emerging technology risk.

The release of the King V Code on Corporate Governance (effective for financial years commencing on or after 1 January 2026) marks a significant shift in how South African boards must approach emerging technologies, particularly generative AI (GenAI).

Why GenAI Pilots Are No Longer "Low Risk"

GenAI systems - whether internal productivity tools, customer-facing chatbots, or content generation engines - introduce dynamic risks including bias amplification, purpose drift, data privacy breaches under POPIA, lack of explainability, and unintended societal impacts.

King V explicitly requires boards to ensure AI deployments adhere to core values: ethics, human centricity, accountability, transparency, explainability, security, privacy, fairness, and trustworthiness, supported by proportionate human oversight and override mechanisms.

Key King V Expectation

The governing body must provide strategic direction for AI and remain accountable for its effective, compliant, and ethical use - including clear human responsibility for decisions, actions, outputs, and outcomes.

10 Critical Questions Boards Must Ask

1

Does this GenAI pilot align with our organisational purpose and King V ethical values?

Boards should demand a clear purpose-alignment assessment. Does the pilot uphold Ubuntu-inspired outcomes of human dignity, fairness, and societal well-being?

2

Who is explicitly accountable for the AI system's decisions, outputs, and potential harms?

King V demands named human accountability. Ask for a RACI matrix that identifies who has override rights.

3

What human oversight and override mechanisms are in place?

High-risk use cases require robust human supervision. Low-risk tools still need defined escalation paths.

4

How will we ensure transparency and explainability to affected individuals and regulators?

In line with POPIA and King V's emphasis on legitimacy.

5

Have we integrated AI-specific risks into our enterprise risk register?

Systemic risks such as bias amplification and model drift must be explicitly assessed.

6

What is the resource and sustainability impact?

In South Africa's energy-constrained environment, boards should evaluate carbon and electricity footprints.

7

How will we measure both financial ROI and non-financial societal value?

King V calls for integrated thinking and ethical-weighted metrics.

8

What assurance processes will be applied before and during deployment?

Independent AI assurance is essential under King V Principle 12.

9

How will we detect and manage purpose drift and shadow AI?

Robust monitoring and drift detection processes must be in place from day one.

10

Are we prepared to say "no" or impose conditions if the risks outweigh the benefits?

Strong governance sometimes means pausing or rejecting a pilot.

From Compliance to Competitive Advantage

Boards that ask these questions proactively demonstrate ethical and effective leadership under King V. They transform AI governance from a compliance burden into a source of stakeholder trust and sustainable value creation.

"In the age of agentic and generative AI, silence from the boardroom is no longer an option. King V calls on governing bodies to lead with courage, curiosity, and conscience."