The European Central Bank (ECB) is considering the potential of large language models, akin to ChatGPT, for tasks such as document analysis and software testing.
While the ECB maintains a cautious stance regarding artificial intelligence, Chief Services Officer Myriam Moufakkir has highlighted an emphasis on addressing data privacy, legal issues, and ethical considerations. In her recent blog post, she stated, “By implementing the right governance, coordination, infrastructure, and investment, we can unify our initiatives on AI and expedite its integration throughout our organization.”
Among the applications being explored are the automated drafting of preliminary code for experts in software analysis and testing. These models also have the capability to analyze, summarize, and contrast documents submitted by the banks under ECB supervision, assisting in the preparation of briefings and summaries.
Additionally, large language models can enhance written communications, making them more accessible to the public. The ECB already employs neural network machine translations to facilitate effective communication with European citizens in their native languages.