The European Central Bank (ECB) is testing how to use artificial intelligence (AI) products to automate some of its tasks. Its “infinity team” is exploring how to generate text, images, and computer software for internal purposes, before considering its use in critical banking operations.

The ECB’s chief services officer, Myriam Moufakkir, has launched an initiative to use AI for a narrow set of use cases. These include drafting briefings, summarizing meetings, creating code for software, adjusting language in official communications, and writing code.

Moufakkir stressed the bank’s cautious approach amid legal and transparency risks. AI often works in a black box manner, making it hard to understand how it arrives at its answers. This raises accountability issues if AI produces harmful outcomes.

Last week, US Federal Reserve governor Lisa Cook said AI without transparency around its decision-making is of little value. A recent analysis suggests that unclear goals for AI in banking reduce its benefits.

The ECB joins other central banks in exploring AI. The Fed is running its FedNow system for instant settlements, which may pave the way for a digital currency. The Bank of England plans to introduce its own system for quick interbank settlements.

However, citizens may not welcome the increased scrutiny introduced by AI. Many people have rejected the idea of a digital currency for its potential to access sensitive transaction data. This concern, along with the use of AI, which has historically been biased against minority groups, may be unacceptable for many in the banked world. The lack of transparency of algorithms with access to unparalleled data could reinforce systemic biases on a global scale.

But Europe may be the best placed to explore this new territory. Its citizens have opposed its digital euro idea as a solution in search of a problem, but the region’s combination of data privacy and AI rules may position it to lead central banks into the next era.

In 2018, the EU passed its General Data Protection Rules legislation, which contained guidelines to protect user data in AI models. Also, Europe’s Artificial Intelligence Act, which could be finalized later this year, will give institutions further guidelines on the ethical handling of user data.

The region has also pioneered rules for digital assets through its Markets in Crypto-Assets bill, which will become effective next year.