Cross‑Jurisdiction AI Governance Paper
A journal article argues for integrated AI governance across the U.S., EU, U.K., Singapore and Hong Kong that treats model risk, consumer protection and operational resilience as a single program. The authors recommend aligning model controls, customer safeguards and continuity planning—an approach framed as especially relevant to regulated financial services. (ijctjournal.org)
A new journal paper says banks should stop governing artificial intelligence in separate silos and run model risk, customer protection, and outage planning as one control program. (ijctjournal.org) The article, published in Volume 13, Issue 2 of the *International Journal of Computer Techniques* and dated April 11, 2026, is by Rajeew Vishvakarma. It compares rules and guidance in the United States, European Union, United Kingdom, Singapore, and Hong Kong. (ijctjournal.org) Its core point is that one bank system can be three things at once: a statistical model, a customer-facing decision tool, and a technology dependency that can fail. The paper says that split oversight lets problems slip between validation teams, conduct teams, and resilience teams. (ijctjournal.org) That framing matches how banks already use artificial intelligence in credit underwriting, fraud detection, collections, trading, customer service, and anti-financial-crime work. The paper proposes one architecture covering inventory, risk tiering, validation, explainability, fairness review, incident response, vendor oversight, and ongoing monitoring. (ijctjournal.org) The timing lines up with a regulatory pileup in financial services. The European Union Artificial Intelligence Act entered into force on August 1, 2024, and the bloc’s Digital Operational Resilience Act has applied since January 17, 2025. (commission.europa.eu) (eur-lex.europa.eu) In the United Kingdom, the Bank of England and Financial Conduct Authority said on November 21, 2024 that 75% of surveyed firms already use artificial intelligence, 33% of use cases rely on third parties, and 55% involve some automated decision-making. The same survey said cybersecurity was the top perceived systemic risk and that only 34% of firms reported “complete understanding” of the systems they use. (bankofengland.co.uk) (fca.org.uk) The United States piece of this stack is older but still active. Federal Reserve guidance known as Supervisory Letter SR 11-7, issued on April 4, 2011, sets expectations for model development, validation, and governance, and the paper treats that as one anchor for bank artificial intelligence controls. (federalreserve.gov) Hong Kong regulators have also linked artificial intelligence to customer treatment. A Hong Kong Monetary Authority circular dated August 19, 2024 addresses consumer protection in the use of generative artificial intelligence and points banks back to its 2019 principles on big data analytics and artificial intelligence. (hkma.gov.hk) The paper does not create new law, and it does not claim the five jurisdictions use identical rules. It says the legal pathways differ, but supervisory expectations are converging around lifecycle controls, explainability, customer safeguards, and resilience when vendors or infrastructure fail. (ijctjournal.org) Its closing recommendation is practical rather than legislative: treat artificial intelligence as an enterprise capability, not a narrow data-science project. For banks already mapping systems to the European Union Artificial Intelligence Act, Digital Operational Resilience Act, and local supervisory guidance, that means one inventory, one escalation path, and one set of controls that can survive an audit and an outage. (ijctjournal.org)
