The International Regulatory Strategy Group report on Artificial Intelligence in financial services identifies a growing global convergence around high-level principles from the OECD, G20 and G7, including human-centricity, transparency, robustness, safety and accountability. The report notes that while these foundational concepts are widely shared, jurisdictions differ significantly in how they translate them into practice, ranging from detailed, prescriptive rulebooks to more flexible, outcomes-focused frameworks and voluntary guidance developed collaboratively with industry.
According to the IRSG’s analysis, Artificial Intelligence is described as a general-purpose technology that can magnify model risk, data governance challenges, third-party concentration risk and cyber threats, particularly for generative Artificial Intelligence, but does not introduce wholly new financial-sector risks. The report therefore argues that supervisory responses should be rooted in existing technology-neutral rules, with interoperable and principles-based oversight rather than the creation of new hard global rules on Artificial Intelligence. The authors warn that given Artificial Intelligence’s rapid evolution, highly rigid international rulebooks risk becoming obsolete and may fail to support innovation.
IRSG Council chair Farmida Bi stresses that as Artificial Intelligence transforms financial services, regulatory strategies must support both innovation and resilience, with coherence but not rigidity, shared taxonomies and supervision through existing frameworks. The report highlights that most jurisdictions draw on OECD, G20 and G7 principles but implement them through different models, citing the European Union’s prescriptive regime, the United Kingdom’s non-statutory, outcomes-focused supervision, and Singapore’s voluntary, co-created guidance. It further cautions that data localisation and extra-territorial measures can fragment markets and impede responsible innovation, and it calls for international cooperation among regulators, policymakers and standard setters to align taxonomies, indicators and supervisory tools so that Artificial Intelligence can be deployed safely and responsibly across borders.
