Find the Right GRC Solution

Search and compare GRC technology built for the problems you’re trying to solve

SFC Chairman Urges Boards to Build Accountable AI Governance, Culture, and Competence

SFC Chairman Urges Boards to Build Accountable AI Governance, Culture, and Competence

By
Regulatory data powered by Corlytics
In a keynote at the Corporate Governance Conference 2026, SFC Chairman Dr.
🔒
Cancel anytime
Full archive access
Custom alerts
Key Takeaways
  • In a keynote at the Corporate Governance Conference 2026, SFC Chairman Dr.
  • Kelvin Wong framed AI oversight as a board-level duty built on three pillars: governance, culture, and competence.
  • The speech creates no new rules, but it signals how the SFC will assess directors' conduct under existing law, including Section 465 of the Companies Ordinance.
  • Compliance teams should treat it as a preview of supervisory expectations and start closing documented AI governance gaps.
Deep Dive

In a keynote at the Corporate Governance Conference 2026, SFC Chairman Dr. Kelvin Wong framed AI oversight as a board-level duty built on three pillars: governance, culture, and competence. The speech creates no new rules, but it signals how the SFC will assess directors' conduct under existing law, including Section 465 of the Companies Ordinance. Compliance teams should treat it as a preview of supervisory expectations and start closing documented AI governance gaps.

The speech sets out the following:
- Governance: Boards should issue board-endorsed, firm-wide AI guidelines covering strategy, capital allocation, risk, and competence, and define where human judgment must not be replaced. It endorses the HKCGI's five priorities: visibility of use, senior ownership, integration into enterprise risk management, human accountability, and transparent reporting. Internal reporting should precede external disclosure and cover material deployments, owners, risk classifications, metrics, incidents, and overdue remediation.
- Culture: Boards should adopt a concise, company-specific AI mission statement that cascades through strategy, procurement, risk appetite, incentives, and assurance.
- Competence: AI literacy should be board-wide, reflected in the board skills matrix, and supported by workshops and scenario exercises.
- Capital and performance: Each material AI investment should state the problem, expected returns, risks, and triggers to scale, pivot, or stop. Dashboards should track value, quality, risk, controls, capability, and trust rather than pilot counts or spend. Incentives should reward early escalation, and "zero incidents" targets are discouraged because they may encourage cover-ups.

The three imperatives are Equip, Align, and Assure.

No deadlines or effective dates were set. The speech references the SFC's June call for licensed firms to strengthen defenses against AI-enabled cyberattacks, including technology inventories, vulnerability management, monitoring, incident response, and recovery. It cites HKCGI research showing nearly nine in 10 issuers mention AI in disclosures, yet fewer than 20% report structured governance frameworks. Risk and incident management, responsible use, and executive ownership are the weakest areas. Smaller issuers and those with concentrated ownership face particular scrutiny, given thin trading and information asymmetry.

Compliance teams should:
- Inventory AI use cases and assign named owners and risk classifications.
- Draft or update board-approved AI guidelines and an AI mission statement.
- Verify that public AI claims match actual controls, to limit disclosure risk.
- Build board reporting that includes incidents and overdue remediation.
- Schedule director training and update skills matrices.
- Add AI-enabled cyber threats to resilience testing.
- Review incentive structures for perverse effects.
- Keep minutes and records showing board challenge and follow-up.

The main exposure is personal director liability and supervisory criticism where AI oversight cannot be evidenced. Disclosure risk arises where AI claims outpace actual governance. Third-party dependence, opacity, cyber risk, and weak accountability are called out as amplified by AI.

Source: Hong Kong Securities and Futures Commission

Oops! Something went wrong