The Algorithmic Commons: Why Central Banks are Betting on Artificial Intelligence
As global markets shift toward high-frequency automated governance, student researchers explore the implications of non-human monetary policy on emerging economies.
Words by
Julian Thorne — London School of Economics
For most of the twentieth century, the operating room of monetary policy was a wood-panelled chamber filled with measured voices. Today, that room is increasingly filled with the soft hum of GPUs.
The Bank of England's quiet 2025 procurement of three forecasting models from a private vendor passed largely without comment in the financial press. Inside the building, it was treated as routine. Outside, it should not have been.
What is genuinely new is not the use of statistics — central banks have run econometric models for decades — but the delegation of inference itself. When a model's outputs cannot be fully audited by the humans nominally responsible for the decision, the chain of democratic accountability frays.
Emerging economies feel this most acutely. A model trained predominantly on OECD data carries the prior assumptions of OECD economies; deployed in Lagos or Jakarta, those priors become invisible policy choices.
The honest answer is that we do not yet know what a fully algorithmic monetary regime would look like in steady state. The honest follow-up is that we are running the experiment anyway.
End of piece
Have a response? Write your own.