The Central Bank of Kenya (CBK) is exploring new ways of using Artificial Intelligence (AI) to monitor economic changes and improve its forecasts, particularly as market conditions become increasingly volatile.

CBK Director of Research Robert Mudida said the regulator has identified nearly 300 potential AI applications across its 14 departments, with some already moving into implementation this year.

However, he maintained that AI should support human decision-making rather than take over critical responsibilities, with the regulator seeking to balance the opportunities offered by automation against the risks associated with its use.

“It is important to emphasise that AI augments intelligence, so it still needs human judgement in a very critical way. We’re not seeing a situation where all main functions will be taken over by AI agents because we still need human judgement to make critical decisions,” he said during an interview.

One of the areas under consideration is the use of AI to complement conventional inflation monitoring by tracking the cost of goods throughout the month, giving researchers additional data to assess emerging trends rather than relying solely on periodic reports.

For example, the director said the technology could help analyse prices on online supermarket platforms by matching products, brands and quantities, in line with the Kenya National Bureau of Statistics' (KNBS) Consumer Price Index (CPI), to calculate inflation.

Another proposed application involves using AI for nowcasting, a technique that estimates real-time economic conditions, by analysing satellite images and converting them into data that can be reviewed alongside CBK’s existing high-frequency economic indicators and estimates before official statistics are released.

This is particularly relevant to GDP data, which is released with a delay of about three months, potentially leaving policymakers without an immediate picture of economic performance during monetary policy meetings.

The bank is also exploring machine learning to improve economic forecasts by identifying errors in previous predictions and adjusting models to account for changing market conditions.

Mudida cited increased volatility following the energy crisis that began in late February as an example of why traditional forecasting methods may require additional support.

The regulator is also examining how AI can analyse large volumes of financial information, including mobile money transactions , to establish patterns and relationships between different economic variables.

The proposed methods are being considered as CBK develops safeguards for the use of emerging technologies in the financial sector.

In September, the regulator published draft Risk Management Guidelines 2026 proposing requirements for institutions using AI, including human oversight, transparency, security, bias testing and regular monitoring to prevent inaccurate or discriminatory outcomes.

The guidelines directed financial institutions to ensure that AI-driven assessments are explainable, supported by reliable data and regularly validated.