European Central Bank Researches Artificial Intelligence for Task Automation Amid Concerns Over Transparency
The European Central Bank is taking baby steps toward integrating AI to automate administrative work.
Using commercially available AI tools, the European Central Bank (ECB) is exploring the possibility of automating routine operations. Before contemplating using it for mission-critical banking processes, the “infinity team” is investigating how to create text, pictures, and computer software.
Myriam Moufakkir, chief services officer at the ECB, has begun a program to use AI in a select group of situations. Some examples of such activities include developing code, generating briefing documents, summarizing meetings, and modifying the language used in official communications.
For Moufakkir, the dangers posed by the law and lack of openness need a careful approach. Accountability for a detrimental response is complicated by the fact that humans cannot always comprehend how AI arrives at its conclusions.
Lisa Cook, governor of the US Federal Reserve, warned this week that AI is useless if its decision-making processes are not made public. Recent research reveals that the advantages of using AI in banking become less obvious when the goal of the technology is poorly defined.
The ECB is the latest central bank to investigate artificial intelligence (AI), joining the FRB, the BoE, and the MAS in Singapore. In the meanwhile, the Bank of England is planning to implement its own system for speedy interbank settlements, while the Federal Reserve is already operating its FedNow system for immediate settlements, which many feel portends a central bank digital currency (CBDC).
It’s unclear how residents will react to the new level of scrutiny imposed by AI. Many people in different nations are wary of a CBDC due to the fact that it may be used to access private financial information.
This worry, combined with the fact that AI has a track record of being cruel to marginalized groups, may be too much for many in the banking elite to bear. The opaque nature of algorithms with access to unprecedented data has the potential to enhance global systemic presumptions.