As financial institutions, consulting firms, and other professional organisations rapidly integrate artificial intelligence into their operations, insufficient human oversight could expose them to serious financial, regulatory, and reputational consequences, according to fintech and compliance expert Opeyemi Kayode.

Kayode made this observation during an appearance on News Central on October 22, 2025, where he discussed the importance of human oversight in preventing artificial intelligence hallucinations, factual errors, and unreliable professional outputs.

Drawing from his experience in finance, internal audit, regulatory compliance, risk management, and process automation, Kayode argued that the central issue is not simply whether artificial intelligence can produce inaccurate information. The greater concern, he said, is whether organisations have established adequate systems for people to verify, challenge, document, and accept responsibility for AI-generated outputs before they are used in business decisions or presented to clients.

According to him, artificial intelligence can generate language, analysis, and conclusions that appear authoritative while containing incorrect information or unsupported claims. Where human professionals fail to review such outputs, the resulting error should not be treated only as a technological failure, but also as a breakdown in human accountability and organisational governance.

“Generative AI can produce information and context that are wrong,” Kayode explained during the interview. “Humans must verify the content, establish its accuracy, and ensure that real people retain ownership and responsibility for the final output.”

He observed that many organisations are adopting artificial intelligence faster than they are developing the safeguards, governance processes, and professional norms required to use it responsibly.

For Kayode, the long-term value of AI should not be measured exclusively by speed or efficiency. Trust must remain a central consideration, particularly in professional services and financial institutions where reports, recommendations, risk assessments, and compliance decisions can directly affect clients, investors, regulators, and members of the public.

“When an organisation releases a report containing false information, it erodes public and client trust,” he said. “Once that trust is damaged, the continuity and credibility of the business may also be placed in jeopardy.”

AI Governance Requires More Than Written Policies

Kayode said incidents involving inaccurate AI-generated reports could accelerate the development of clearer governance frameworks within professional services firms and among regulators.

However, he cautioned that effective governance cannot be achieved merely by drafting an artificial intelligence policy or requiring an employee to conduct a superficial review. Instead, organisations need a multilayered oversight structure that addresses internal accountability, independent verification, and regulatory compliance.

The first level, he explained, should involve internal policies that clearly identify who is responsible for reviewing AI-generated material. Such policies should also require explainability and documentation throughout the lifecycle of an AI-assisted project, analysis, presentation, or report.

This means organisations should be able to explain where artificial intelligence was used, what information was provided to the system, how the output was validated, and which human professional approved the final result.

The second level should involve independent audits and verification mechanisms. These controls would help organisations assess the integrity of the data used by AI systems, trace the sources supporting generated content, and determine whether outputs comply with professional and industry standards.

The third level should be supported by external regulatory frameworks. Kayode referred to emerging standards such as the European Union Artificial Intelligence Act and the Artificial Intelligence Risk Management Framework developed by the United States National Institute of Standards and Technology.

He explained that these frameworks reflect a growing expectation that artificial intelligence should not operate as an unexplained “black box,” especially when its outputs affect customers, employees, borrowers, investors, or regulated institutions.

Organisations using AI should therefore be capable of documenting how the technology influenced a decision and explaining the processes used to validate its conclusions.

Lessons for Banks and Fintech Companies

The consequences of inadequate AI oversight may be particularly serious in banking and financial technology, where artificial intelligence is increasingly used for credit risk assessments, fraud detection, transaction monitoring, regulatory reporting, customer verification, and lending decisions.

Kayode acknowledged that AI allows financial institutions to analyse significantly larger volumes of information than human employees could process independently. Nevertheless, he warned that these applications involve high-stakes decisions in which even a relatively minor error could lead to substantial financial losses, regulatory violations, reputational damage, or unfair treatment of customers.

“This is a major wake-up call for banks, fintech companies, and other adopters of AI,” he said. “Artificial intelligence cannot be left unchecked. Human involvement is not negotiable.”

He recommended maintaining a meaningful “human in the loop” process in which qualified professionals do more than merely observe automated decisions. The human reviewer should verify the accuracy of the output, challenge unsupported conclusions, assess the underlying information, and assume responsibility for the final decision.

Kayode also highlighted the risk that existing human biases may be transferred into artificial intelligence systems. In lending, for example, historical decisions made by people may contain patterns of unfairness or discrimination. Where those historical records are used to develop automated models, the resulting system may reproduce or amplify the same biases.

For this reason, he said financial institutions need secondary review processes, independent testing, data integrity controls, and identifiable professionals who can be held accountable for AI-assisted decisions.

Balancing Innovation With Accountability

Kayode’s comments reflect a broader challenge confronting financial institutions and professional services firms. Organisations are under pressure to adopt artificial intelligence to improve productivity, reduce costs, identify risks, and respond more quickly to customers. At the same time, the use of AI creates new responsibilities relating to accuracy, transparency, governance, and accountability.

For organisations operating in highly regulated sectors, the strongest approach may therefore be one that combines technological innovation with the principles already familiar to financial control and risk professionals: segregation of duties, independent review, documented evidence, clear ownership, escalation procedures, and continuous monitoring.

Kayode maintained that artificial intelligence should support professional judgment rather than displace responsibility for it. The organisations that benefit most from AI, he suggested, will not necessarily be those that adopt it fastest, but those that establish systems capable of protecting the accuracy, fairness, and credibility of its outputs.

About Opeyemi Kayode

Opeyemi Kayode is a finance, technology, risk management, and regulatory compliance professional currently serving as Manager, Special Projects, Consumer Solutions at Solenis Inc. in Wilmington, Delaware, where he leads strategic financial analysis and business initiatives supporting major corporate investments and growth opportunities.

Before joining Solenis, Kayode built his career at KPMG in the United Kingdom and Nigeria, where he advised financial institutions and other regulated organizations on internal audit, financial controls, governance, regulatory compliance, and enterprise risk management. His experience across the Big Four helped establish him as a recognized financial expert and emerging fintech thought leader, with particular expertise in the responsible adoption of artificial intelligence, financial governance, and technology-driven risk management.