AI in finance will advance only as fast as regulators allow
Dubai's virtual assets chief says supervisory infrastructure matters as much as the rules themselves—and jurisdictions that lag will be bypassed.
Financial regulators occupy a unique position in the AI revolution: they are simultaneously being transformed by artificial intelligence and serving as gatekeepers that determine how much of its benefits reach the financial industry. That dual role will shape the next decade of finance more than any single technology or policy, according to Matthew White, CEO of Dubai's Virtual Assets Regulatory Authority.
White, who spent over a decade advising financial institutions on cybersecurity and technology risk before building VARA from the ground up, argues that the infrastructure through which oversight is delivered matters as much as the rules themselves. His perspective comes from leading the world's first dedicated virtual assets regulator, where traditional supervisory models have proven inadequate.
The regulatory gap
The starting point across financial regulation is uneven. While tier-one banks have spent billions on regulatory technology over the past decade, Thomson Reuters' annual Cost of Compliance surveys show compliance budgets and headcount climbing steadily with productivity gains that are linear at best. The 2023 wave of large-language-model adoption produced widespread experimentation but minimal production deployment in compliance-critical workflows—the cost of a wrong answer in regulation remains higher than the cost of a slow one.
Among regulators, the gap is wider still. Authorities like the Monetary Authority of Singapore, the UK's Financial Conduct Authority, and the Bank of England run serious supervisory-technology programs. Many others operate predominantly on PDF rulebooks, sample-based onsite inspections, and email-driven supervisory cycles.
Virtual assets as proving ground
Virtual assets matter beyond their own market because they impose different constraints on supervision. Assets that are programmable, operate around the clock, remain cryptographically auditable, and ignore borders cannot be supervised through quarterly inspections and paperwork. Regulators in this domain must build genuinely new capabilities: on-chain audits rather than sampled reviews, real-time monitoring rather than after-the-fact reporting, and programmable compliance rather than rule-engine workflows.
The boundary between virtual assets and traditional finance is dissolving. McKinsey's 2024 analysis projects $2 trillion in tokenized financial assets by 2030. The Bank for International Settlements' Project Agorá, now in testing with seven central banks and over forty financial institutions, is bringing tokenized wholesale settlement into live infrastructure. The supervisory technology being built for virtual assets today will run mainstream finance within a decade, White contends.
Two waves of transformation
White identifies two distinct waves of AI transformation in regulation. The near-term wave, arriving over the next 12 to 24 months, will compress authorization reviews from weeks to days and reduce false-positive rates in compliance alerts—currently between 90% and 95% according to industry research.
The medium-term wave, arriving in three to five years, will fundamentally break the current model. The interface between regulators and licensees will shift from periodic submissions to continuous data exchange. Once data flows in real time, the operational case for sample-based inspection collapses.
Why it matters
AI advances in finance only as fast as regulators permit. A firm can build supervisory-grade compliance infrastructure, but if its regulator still demands PDF returns and human-mediated supervisory cycles, productivity gains remain trapped inside the firm. Capital follows the regulators that move—jurisdictions that fall behind get bypassed. The legal architecture of compliance assumes human decision-making throughout, and until regulators resolve questions of accountability when AI generates suspicious activity reports or contributes to fitness-and-propriety judgments, firms will hold back the highest-impact applications. The bottleneck is regulatory rather than commercial.
Regulators who treat the AI transition as something happening to their industry will become the bottleneck first and the casualty second, White warns. Those who treat themselves as the engine will inherit the operating environment of finance.
These details were first reported by Fortune in a commentary piece by White.
This is an original analysis by the Omega editorial team. Source reporting: AI Watch.
Want systems like this working for your business?
Book a Call
