AI Is Challenge for Central Banks to Raise Their Game, BIS Says

Bastian Benrath, Bloomberg News

Credit: BIS ,

(Bloomberg) -- Central banks must get a better grip on artificial intelligence both to gauge its economic impact and to harness the technology for themselves, the Bank for International Settlements said.

In a report published on Tuesday, Basel-based officials said that AI won’t supplant human judgment but that its likely effects on productivity, investment and consumption mean policymakers must find ways to study that and employ it for their own benefit.

“The rapid and widespread adoption of AI implies that there is an urgent need for central banks to raise their game,” the BIS said. “Central banks need to upgrade their capabilities both as informed observers of the effects of technological advancements as well as users of the technology itself.”

Among the economic effects listed are productivity, with the example that software developers can code more than twice as many projects per week if they are supported by AI. Meanwhile companies will likely invest more in information technology, while for consumers, algorithms can yield better choices in spending decisions, thereby boosting consumption.

For central banks, the ability of new algorithms to crunch vast amounts of data also holds great potential for tasks like nowcasting and forecasting, according to Cecilia Skingsley, head of the BIS Innovation Hub, who told reporters that “it will lead to better decision-making.”

Economic Adviser and Head of Research Hyun Song Shin added that setting interest rates will probably remain the task of humans because that involves judgment. Skingsley emphasized that an algorithm can’t answer for its own actions.

“The ways we organize ourselves and our societies are that we like to hold human beings accountable to decision-making,” she said. “You know, changing politicians, possibly changing central-bank governors from time to time. So I can’t really see a future where it will be an AI setting rates.”

The new technology’s effect on inflation will depend on expectations of welfare gains and will differ in the short and the long run, the BIS said. As long as consumers underestimate future benefits, such new capabilities will dampen consumer-price growth, officials said, noting that such an effect may reverse when people start to anticipate the advantages.

The rise of AI will also expose central banks to risks including hacker attacks, disinformation, and unknown biases, according to the report. Some of those are already a worry, officials have said.

If a majority of market participants rely on the same set of algorithms from dominant cloud computing firms, this could pose dangers to financial stability and even result in unintended collusive behavior, officials said.

For financial institutions, AI poses similar opportunities and risks, particularly in payments, insurance, lending and asset management, the BIS said. Opportunities range from gains in risk assessment and portfolio allocation to liquidity management. Potential pitfalls include the impact of cyber attacks, herd behavior of algorithms and privacy concerns.

To mitigate such risks and harness the technology’s potential, the report called on monetary authorities to share experiences, tools and data.

“There is an urgent need for central banks to collaborate in fostering the development of a community of practice,” the BIS said.

--With assistance from Jeff Black.