AI and Quantum Computing:

Central banks and financial authorities face two technology transformations moving on very different clocks. Artificial intelligence is already changing how institutions analyse data, organise work, supervise firms and defend critical systems. Quantum computing remains experimental, yet the migration required to protect cryptographic trust must begin well before a sufficiently capable machine exists.

The common challenge is not predicting the technology perfectly, but deciding how a public institution should act before certainty while preserving judgement, accountability, independence and public trust. For AI, the immediate task is to move from fragmented experimentation to measurable value at scale. For quantum computing, near-term computational value remains uncertain, but the security transition is already actionable.

The roundtable therefore advances a two-speed agenda: scale AI deliberately by prioritising high-value applications and embedding risk, ethics and human accountability from the outset; and prepare for quantum early by identifying cryptographic exposure, building crypto-agility, protecting long-lived information and coordinating migration across the financial ecosystem.

Across both technologies, the discussion will examine how central banks can avoid both hype and paralysis: acting early where delay compounds risk while resisting premature technology bets that are not linked to mandate value.

Guiding Questions

• Where can AI materially improve policy outcomes, resilience or institutional effectiveness, and how should that value be measured?
• How should supervisory and financial-stability frameworks evolve as AI becomes more autonomous, concentrated and systemically embedded?
• What no-regret quantum-readiness actions should central banks initiate now, and what should remain exploratory?
• Which capabilities must be owned internally, which can be shared, and where can external dependence be accepted without weakening institutional agency?
• How can central banks build scarce technology capability while retaining vendor neutrality, accountability and strategic optionality?