The European Central Bank (ECB) is considering the use of large language models, similar to ChatGPT, for various purposes such as document analysis and software testing.
The ECB is approaching the use of artificial intelligence (AI) cautiously, taking into account concerns related to data privacy, legal constraints, and ethical considerations, as stated by Myriam Moufakkir, the Chief Services Officer at the ECB, in a blog post.
Moufakkir also emphasized that by establishing proper governance, coordination, infrastructure, and investment, the ECB aims to streamline its efforts related to AI and accelerate its adoption within the organization.
The ECB is exploring several potential applications for large language models, including generating initial drafts of code for experts to use in analysis and software testing. These models can also assist in analyzing, summarizing, and comparing documents prepared by the banks that the ECB supervises, as well as help in preparing summaries and draft briefings.
Additionally, large language models can contribute to improving the clarity and comprehensibility of the ECB’s communications, making them more accessible to the public. The ECB is already using neural network machine translations to communicate with European citizens in their native languages.
