CBK: Explores AI Agents Banking Operations, Prioritizes Oversight

• The Central Bank of Kenya is exploring the use of AI agents to automate certain banking operations, including payments.
• CBK Director of Research Robert Mudida stated that human oversight will be maintained for AI-assisted decisions.
• Agentic AI systems can independently plan and execute multi-step tasks, presenting both significant opportunities and dangers.
• The CBK is also evaluating over 300 other AI applications and testing an AI-powered chatbot to improve efficiency.
• Data security, governance, and regulatory compliance are key considerations as the CBK weighs AI's benefits against its risks.

Robert Mudida, the Director of Research at the CBK, highlighted that while the institution is exploring these advanced applications, a critical principle of human oversight will be maintained for all decisions made with the assistance of this technology.

The Central Bank of Kenya (CBK) is actively investigating the integration of artificial intelligence agents into its banking operations, aiming to streamline processes while diligently managing associated risks. This strategic move by the regulator seeks to leverage cutting-edge technology to automate various functions, with particular attention given to areas such as payments.

Robert Mudida, the Director of Research at the CBK, highlighted that while the institution is exploring these advanced applications, a critical principle of human oversight will be maintained for all decisions made with the assistance of this technology. He specifically referenced agentic AI, defining it as sophisticated software systems capable of independently planning and executing multi-step tasks using digital tools, rather than merely responding to individual prompts. This distinction underscores the advanced nature of the AI systems under consideration by the Central Bank Kenya artificial intelligence initiatives.

Mudida emphasized the significant potential of agentic AI, noting its capacity to undertake critical decisions traditionally performed by human personnel. He described this form of AI as being at the “highest level” of technological sophistication, presenting both substantial opportunities and considerable dangers. Despite the inherent risks, the CBK is proceeding cautiously, recognizing the specific advantages that agentic AI could offer in sectors like payments.

Beyond agentic AI, the CBK is also evaluating more than 300 distinct artificial intelligence applications designed to enhance overall operational efficiency. This broad exploration includes the testing of a new AI-powered chatbot. Mudida affirmed the CBK's commitment to rapid technological adoption while simultaneously prioritizing robust AI governance to safeguard data security, a crucial aspect of Kenya banking AI governance.

The CBK's Research Department is already utilizing artificial intelligence in practical applications, employing it for microforecasting in dynamic economic environments. Additionally, AI is used for nowcasting, a method to bridge data gaps and furnish the Monetary Policy Committee with economic information in near real-time. This existing deployment demonstrates the CBK's foundational experience with AI, setting the stage for its deeper exploration into agentic systems.

The regulator's interest in CBK explores AI agents banking operations aligns with a broader trend across the financial sector, where institutions are increasingly adopting AI for functions such as payment processing, fraud detection, credit risk assessment, and customer service. Robert Mudida CBK AI statements confirm that the institution is carefully balancing the potential benefits of this technology against critical concerns, with data security, governance, and regulatory compliance remaining paramount considerations for the future of AI in Kenyan financial sector.

Lawyers and compliance officers in Kenya's financial sector should closely monitor the Central Bank of Kenya's (CBK) exploration of AI agents, particularly regarding future regulatory frameworks or guidelines for their deployment. This includes anticipating new compliance requirements related to data security, AI governance, and accountability for decisions made by agentic AI systems, to advise clients effectively on emerging legal risks and ensure operational compliance.

Source: Original reporting via Capital FM