How much do organisations need to understand about the AI systems they deploy? Enterprise AI can create a “visibility gap” where organisations remain accountable for the outputs generated by the AI even when they cannot fully inspect or explain the underlying technology.
The goal is not perfect technical transparency, but sufficient visibility to govern AI responsibly. Drawing on regulatory developments in Hong Kong, Mainland China, the EU and the US, this article explores how explainability becomes a practical governance and contracting issue. Effective oversight, access to relevant documentation and records, supplier cooperation with investigations, and the ability to review, challenge or override AI-generated outputs all help bridge that gap. The key question is whether the organisation has the information, rights and controls needed to explain, investigate and remain accountable for how AI is used.
Read the full article here.
Introduction
In its article of 5 November 2026, the Insurance Authority (IA) reports that premium financing accounted for 36% of new business premiums in the first half of 2026, compared with around 12% in 2024. The IA highlights three trends which industry players should note.
Premium financing involves borrowing to pay insurance premiums and assigning policy rights to the lender as security. It adds borrowing risks to the insurance purchase, which should not be overlooked when assessing the potential returns.
IA’s key observations
Foreign-currency loans: There have been arrangements under which offshore banks offer loans in low-interest-rate currencies, such as the Swiss franc, for premium financing. Such arrangements present interest-rate and exchange-rate risks. Overseas loan arrangements may also make complaints and legal proceedings more difficult and costly.
Wider retail participation: The IA cautions against lowering entry thresholds to attract middle-class customers. Leveraged borrowing is generally suitable only for high-net-worth individuals with sufficient liquid assets, stable income and high risk tolerance.
Higher policy leverage: Premium discounts, higher Day 1 surrender values and generous lending terms can combine to produce high leverage. A small upfront payment should not be mistaken for a lower affordability threshold.
Regulatory developments
The IA confirms that its third joint inspection with the Hong Kong Monetary Authority (HKMA) is underway. Further regulatory measures will be considered in light of the findings.
Separately, the HKMA’s 19 November 2025 circular, applicable to new facilities from 1 January 2026, says banks should not promote premium financing as a strategy to enhance returns. Specified standards also apply to institutions and subsidiaries acting solely as lenders.
Practical tips
Insurers: review product design, lapse assumptions and liquidity risk against actual financing concentrations and leverage.
Intermediaries: assess the full funding arrangement and explain how lower policy benefits, higher interest rates and adverse exchange rates may affect customers simultaneously.
Banks and lenders: review loan tenors, repayment terms, interest-rate disclosures and sales scripts, taking account of their role in each arrangement.
For more information, see the IA’s thematic webpage on premium financing.
How the Hong Kong Court of First Instance dealt with a Canadian bank’s argument that enforcing a EUR 30 million arbitral award would expose it to a real risk of criminal prosecution under Canadian sanctions law.
In A Company v The Bank [2026] HKCFI 3169, the Court of First Instance considered whether enforcing an arbitral award would contravene Hong Kong public policy. A Canadian bank (the “Bank“) argued that enforcement would expose it and its employees to a real risk of criminal prosecution under Canadian sanctions law. The Court dismissed the Bank’s application to set aside the enforcement order, holding that the Bank had not established a real risk of prosecution; even if established, enforcement would not have contravened Hong Kong public policy.
Background and Canadian Sanctions
The Arbitration and GAC’s Response
The Enforcement Proceedings
The Court’s Reasoning
Hong Kong Public Policy
The Exchange published a Consultation Paper on Listing Framework Competitiveness Review (Phase 2) on 21 September 2026, seeking market feedback on reforms to the post-listing regulatory framework. The proposals aim to ensure that the framework remains fit for purpose, while enhancing Hong Kong’s competitiveness and attractiveness as an international listing venue.
Following its Phase 1 review of initial listing requirements, the Exchange’s Phase 2 review addresses the continuing requirements applicable to issuers after listing. In particular, the Consultation Paper proposes reforms to (1) notifiable transactions, (2) connected transactions and (3) spin-off transactions to give issuers greater flexibility while maintaining investor protection.
Read the full article here.
“AI contracts often function as the mechanism through which governance expectations are converted into operational obligations.”
“AI procurement is no longer simply about acquiring technology. It is increasingly about preserving strategic flexibility and avoiding excessive dependence on any single provider.”
As enterprise AI moves from pilot projects to business-critical infrastructure, contracts governing its procurement are becoming the primary mechanism for AI governance. Boards, senior management and procurement teams traditionally often place focus on data ownership when AI services are procured from third parties, but data ownership is only the starting point. The more significant questions concern how the information supplied may be used after the end of the contractual arrangement, who controls the resulting capability if the relationship between the organisations providing the information and the AI service providers changes, and who remains accountable when an AI system produces unexpected outcomes.
This article examines how enterprise AI contracts allocate value, control and accountability across the parties involved, and sets out a practical framework for reviewing existing or proposed AI arrangements before problems arise.
Read the full article here.
See news from our global offices