The European Central Bank (ECB) is considering using artificial intelligence (AI) to better understand inflation, after years of underestimating price pressures and late implementing the current monetary tightening cycle, now its most aggressive. of its history.
Like many companies already using AI, the ECB is looking for ways to process and analyze millions of pieces of data, including public price information, company statistics, press articles and banking supervision documents, in order to achieve better analysis for its monetary policy decisions.
“AI gives us new ways to collect, clean, analyze and interpret this mass of available data, so that the information collected can drive work in areas such as statistics, risk management, banking supervision and asset analysis. monetary policy”, can be read in a Message published by the ECB on Thursday in a blog.
AI to simplify the ECB’s language
While the ECB has underestimated inflation for years, some bank officials have openly questioned its models and their viability, which is based on figures that require constant upward revisions.
Among the various AI initiatives, the bank wants to deepen its understanding of price behavior and inflation dynamics, according to the blog post. Using web scraping, the ECB can collect large amounts of price data in real time, but the figures are unstructured and not suitable for calculating inflation. The ECB therefore wants to exploit AI to structure data and improve its analysis.
Another initiative involves automating the data classification process of tens of millions of companies, banks and public sector entities, in order to better understand their financial situation. The ECB also hopes to use AI to simplify its communication, to respond to critics who consider its language too complex and technical, impossible for ordinary people to understand.
“A developed language model can also help improve texts written by ECB officials, making their communication easier to understand for the public,” writes the central bank. “On a related note, we have been using machine translation neural networks for some time to help us communicate with European citizens in their native language,” the bank adds.
Source: BFM TV
