The European Central Bank is using artificial intelligence (AI) to perform some tasks and exploring other possible uses in the future, according to a blog published today.
The blog is written by the ECB's chief services officer Myriam Moufakkir.
It outlines how AI offers new ways for the bank to "collect, clean, analyse and interpret" data used in its work in statistics, risk management, banking supervision and monetary policy analysis.
AI is already being used to streamline its collection of statistics from various institutions in euro member states. This data comes from over "ten million legal entities" across Europe.
It describes manually collecting and collating this data as "very time consuming" so it is employing machine-learning techniques to automate the process.
It is applying "web scraping and machine learning" to collect information on the individual prices of goods and services. Much of this data is "unstructured" and therefore unsuitable for measuring inflation, so the ECB is using AI to help with this.
The bank is also using AI in banking supervision and has created a platform called "Athena" to collect documents for use by regulators.
The ECB is now examining the possible uses of large-language models like ChatGPT.
One of the possible uses might be to "more quickly prepare summaries and draft briefings" for its staff.
It also says it could be used to "help improve texts being written by staff members" and making the ECB's communication "easier to understand for the public".
The blog details how the ECB already uses "neural network machine translations" to help them "communicate with European citizens in their mother tongues".
