Nigerian researcher patents AI compliance system amid U.S. finance risk warnings
A Nigerian artificial intelligence researcher based in the United States has secured a national patent for a system designed to automate regulatory compliance within AI-powered investment platforms, as American financial regulators intensify scrutiny of algorithmic risks in capital markets.
A Nigerian artificial intelligence researcher based in the United States has secured a national patent for a system designed to automate regulatory compliance within AI-powered investment platforms, as American financial regulators intensify scrutiny of algorithmic risks in capital markets.
Ayobami Olanrewaju, a Lagos-born AI researcher, received Nigerian Patent NG/PT/NC/O/2026/21968 in March 2026 for what is described as an ‘AI-driven adaptive portfolio optimisation system using deep reinforcement learning for dynamic asset allocation in volatile financial markets.’
He said his patented system was developed to address that gap. “Most current AI investment systems treat compliance as something you bolt on after the model decides.
“What we have done is move compliance inside the learning process itself, so every trade decision the system produces already carries an auditable record of how regulatory risk limits were enforced.”
The patent outlines eight claims focused on how machine-learning systems can make investment decisions while remaining transparent and auditable for regulators.
The development comes amid growing concerns among U.S. regulators about the increasing use of artificial intelligence in finance.
In its 2023 annual report, the U.S. Financial Stability Oversight Council (FSOC) identified AI as a potential vulnerability to the American financial system, warning that rapid advances in artificial intelligence could introduce systemic risks if left unchecked.
Similarly, the U.S. Consumer Financial Protection Bureau (CFPB) stated in Circular 2023-03 that financial institutions using algorithms are still fully accountable under existing consumer protection and fairness laws.
A January 2025 report by the U.S. Government Accountability Office (GAO-25-107197) also highlighted a shortage of specialists capable of embedding compliance safeguards directly into AI-driven financial systems.
The system uses deep reinforcement learning, a branch of artificial intelligence where algorithms learn through continuous interaction with market data, to dynamically adjust asset allocations across stocks, bonds, and commodities.
Unlike traditional portfolio models that rebalance investments at fixed intervals, the patented framework applies value-at-risk thresholds in real time, integrates automated compliance checks into every trading cycle, and factors transaction costs directly into allocation decisions.
According to Olanrewaju, the approach addresses a major limitation in many academic investment models that often perform well in simulations but struggle once real-world trading costs and regulatory requirements are introduced.
Backtesting against the Standard & Poor’s 500 Index reportedly showed a five per cent improvement in risk-adjusted returns compared to traditional strategies.
The model also demonstrated the ability to shift investments into lower-volatility assets during market downturns before returning to growth-focused assets as conditions stabilised.
Research linked to the patent has been published in the International Journal of Research in Finance and Management and later expanded in peer-reviewed journals including Finance Research Letters, Expert Systems with Applications, Equilibrium Quarterly Journal of Economics and Economic Policy, and Nature Scientific Reports.
Subsequent studies have extended the framework into related areas such as explainable artificial intelligence for credit decisions and dynamic risk prediction in financial-production systems.
“The U.S. financial system has a workforce-shortage problem in exactly this niche where people who can build AI systems that regulators can actually audit,” Olanrewaju said. “There is no reason that gap cannot be closed by researchers trained in our part of the world.”
He also disclosed plans to file a U.S. provisional patent later this year for a narrower component of the system focused on explainable AI compliance monitoring using SHAP-based attribution technology, which allows regulators and portfolio managers to trace the factors behind algorithmic investment decisions in real time.
According to him, the filing is expected to support pilot deployments with U.S.-based portfolio managers in 2027.
“The provisional filing in the United States this year is the next step. We move from patent on paper to pilot deployment with regulated portfolio managers in 2027,” he said.
The development places Olanrewaju among a growing group of researchers working on governance-focused artificial intelligence systems for highly regulated financial environments.
The U.S. National Institute of Standards and Technology’s AI Risk Management Framework released in 2023 called for trustworthy by design AI systems, though practical commercial applications in fast-moving portfolio management remain limited.
“I want Nigerian innovation visible inside U.S. financial regulation, not just inside U.S. financial returns,” Olanrewaju said. “If a system built by a Lagos-born researcher becomes part of how U.S. regulators verify that AI in capital markets is safe, that is the kind of contribution our generation should be making.”
Olanrewaju is a fellow chartered economist of the Institute of Chartered Economists of Nigeria and has also served as a judge for the U.S. Army-supported eCYBERMISSION national STEM competition.
His published research focuses on explainable artificial intelligence and risk governance in U.S. capital markets, with more than 110 citations across peer-reviewed publications.
