In a keynote speech, European Central Bank Executive Board member Philip R. Lane summarised ECB staff work on how artificial intelligence is affecting the euro area economy, financing conditions and banking, while stressing that the aggregate implications for productivity, employment and inflation are still uncertain at this stage. The speech also set out how the ECB is expanding its own use of AI in analysis and operations and intends to move from pilots to systematic implementation over 2026 and 2027. ECB analysis cited rapid diffusion among workers, with Consumer Expectations Survey data showing AI use rising from 26% of employees in 2024 to 40% in 2025, alongside marked differences by age and education. Digital investment has risen strongly since 2014 but remains well below the United States, with a proxy suggesting it accounted for about 12.4% of total investment in the euro area versus 24.3% in the United States in 2024. Scenario work presented in the speech linked faster diffusion to larger total factor productivity gains over the coming decade, around 0.3-0.4 percentage points per year, versus roughly 0.2 percentage points per year under slower adoption. On employment, the evidence was described as inconclusive, with little sign so far of a substantial aggregate impact, even as large firms contacted in the Corporate Telephone Survey reported AI-enabled process optimisation as a factor in a subdued employment outlook. Lane also highlighted financing frictions for AI investment in the euro area, including shallow venture capital and private credit markets, and noted that high AI-active sectors have recently seen moderating bank loan growth while market-based funding has increased, with debt securities issuance growth reaching 13% in January 2026. For banks, supervisory data cited more than €4 billion of digital-technology investment in 2025 and AI use at nearly 90% of significant institutions, alongside risks ranging from cyber threats and operational dependencies on concentrated providers to spillovers from any abrupt repricing of frontier AI valuations. Monetary policy implications were presented as highly uncertain, and the latest staff estimates of the euro area equilibrium rate did not indicate a material shift. Internally, the ECB reported expanded use of machine-learning tools for forecasting and nowcasting, including inflation density forecasts using quantile regression forests, and a generative-AI tool that cuts Corporate Telephone Survey write-up time from about an hour to 20-30 minutes with human review. Over 2026 and 2027, the ECB plans a central banking digitalisation programme built around a data analysis hub, an AI research lab and an “assistants studio” for deploying AI assistants within a controlled environment aligned with governance safeguards and the EU AI Act.

Original source View in tracker