As global economic and political volatility intensifies with the ongoing Iran conflict and other geopolitical tensions, the European Central Bank (ECB) is deploying artificial intelligence to sharpen its understanding of inflation risks. In a detailed analysis published on 21 April 2026, senior ECB economists highlight how a machine learning model now provides experts with timely insights into the chances that inflation could surge or fall well beyond official projections.
The ECB explained that this innovation addresses a core challenge for policymakers: in today’s fast-changing environment, price pressures can shift rapidly, making it essential to look beyond a single “baseline” forecast and evaluate the full spectrum of possible outcomes.
Traditional economic models used by the Eurosystem rely on a limited set of indicators and often impose strict assumptions that overlook complex interactions.
The new tool, known as a quantile regression forest (QRF), overcomes these limitations.
It draws on an extensive dataset of around 60 variables covering inflation expectations, cost pressures, real economic activity, and financial conditions—information that ECB staff already monitor closely.
Unlike conventional approaches, the QRF captures non-linear relationships and intricate data patterns, delivering not only point forecasts but also a complete picture of upside and downside risks.
The ECB began integrating the QRF into its regular analytical toolkit at the end of 2022, initially for short-term inflation forecasting and risk assessment. It has since been extended to nowcasting GDP growth. During 2025, the model proved particularly valuable.
It identified wages and firms’ selling-price expectations as the primary drivers behind revisions to core inflation projections (excluding energy and food).
By scanning real-time data, the QRF flagged emerging risks across different components of the headline inflation measure (HICP) even when traditional indicators sent mixed signals in the volatile post-pandemic period.
A practical illustration comes from the ECB’s quarterly projections for 2025.
When the QRF’s uncertainty bands placed official forecasts in the lower range, actual core inflation later exceeded projections by about 20 basis points—confirming the model’s early warning of upside risks.
As each quarter progressed, the QRF’s probability ranges narrowed and reliably encompassed the final outcomes, demonstrating growing precision.
In contrast, baseline projections sometimes sat outside these bands, underscoring the added value of the AI-driven risk lens.
Beyond forecasting accuracy, the QRF offers clear economic interpretations. It can pinpoint which factors—such as wage growth, import costs, or expectations—are pushing risks higher or lower at any moment.
This interpretability helps policymakers understand not just what might happen, but why.
ECB experts see machine-learning tools like the QRF playing an expanding role.
They complement rather than replace conventional models, handling surging data volumes efficiently and revealing sector-specific dynamics or non-linear effects that have become more prominent in recent years.
In an environment of persistent uncertainty, these capabilities will strengthen the Eurosystem’s ability to monitor trends, refine forecasts, and inform monetary policy decisions with greater agility and depth.
By embracing AI, the ECB like other international entities is effectively equipping itself to navigate turbulent times more effectively—ensuring that inflation risks are no longer hidden in the fog of uncertainty but brought into sharper focus for well-calibrated responses.