The financial stability implications of artificial intelligence and digital finance
Remarks by Mr Tao Zhang, BIS Chief Representative for Asia and the Pacific, at International Financial Week, in conjunction with the Asian Financial Forum (AFF), Hong Kong, 26 January 2026.
Remarks prepared for delivery.
Introduction
Thank you for the invitation to speak at this important forum. It is a pleasure to join you here in Hong Kong.
The title of this session is "Development and innovation of AI and digital finance". These changes are reshaping how financial services are delivered, how markets function and how risks are managed. From a central bank perspective, they also raise important questions for financial stability.
Today, I would like to focus on the financial stability implications, including tokenisation, which is a central element in current policy discussions.
My remarks will proceed in three parts:
- First, I will briefly discuss how artificial intelligence (AI) and digital finance are developing and why they matter for the functioning of the financial system.
- Second – and this will be the core of my remarks – I will turn to the financial stability implications, drawing on recent analytical work by the Bank for International Settlements (BIS) and the Financial Stability Board (FSB).
- Finally, I will say a few words about international cooperation and the role of the BIS.
Development and innovation: implications for market functioning
Let me begin with how AI and digital finance are developing, and why they are attracting such attention from policymakers.
AI is being adopted across the financial sector for a wide range of purposes. Financial institutions use AI to process large volumes of data, support credit underwriting, detect fraud, manage risks and automate back-office functions. More recently, advances in large language models and generative AI have expanded the range of possible applications, including customer interaction, internal analysis and supervisory processes.
The drivers of AI adoption are well understood. On the supply side, rapid advances in computing power, data availability and model capabilities have lowered barriers to entry. On the demand side, firms are seeking productivity gains, cost reductions and competitive advantages, while authorities are exploring the use of AI to enhance regulatory and supervisory effectiveness.
Digital finance, more broadly, refers to the increasing digitalisation of financial assets, processes and infrastructures. A key component of this is tokenisation which, loosely speaking, is the representation of financial assets, such as securities or deposits, in digital form using technologies such as distributed ledger technology.
As we have already witnessed, tokenisation affects how financial transactions are organised and executed. It can bring trading, settlement and collateral management closer together, reduce reconciliation costs and support more efficient use of liquidity and collateral. Tokenisation may also reduce frictions in cross-border payments and securities settlement – an issue of particular relevance for regions with deep trade and financial linkages, including Asia.
Taken together, AI and digital finance can improve efficiency, reduce costs and support more integrated financial markets. However, these same developments also change the way risks arise and propagate across the financial system, and they post challenges for regulators and supervisors. In short, they have strong financial stability implications.
Financial stability implications: three channels
Let me now turn to the channels through which the financial stability risks may rise. Recent research and policy analysis suggest that AI and digital finance may affect financial stability through multiple channels. To illustrate, I will focus on the three channels that are particularly relevant from a central bank perspective, because they relate to market functioning, operational resilience and the propagation of stress through the financial system.
Market functioning and liquidity
First, AI and digital finance can affect market functioning and liquidity:
- AI can speed up trading and portfolio adjustments, which may intensify short-term price movements when market conditions change.
- Tokenisation represents financial claims digitally, which can make transactions and settlement more efficient. But digital claims may move or be redeemed more quickly than the underlying assets can be sold or funded.
- In normal conditions, these features can improve efficiency. But in periods of stress, faster trading and faster-moving claims can strain liquidity, amplify volatility and contribute to disorderly market conditions.
Operational dependencies and resilience issues
Second, these technologies raise operational risk and resilience issues:
- AI systems often depend on specialised hardware, cloud computing services, external data providers and pretrained models, many of which are concentrated among a small number of providers. Digital finance and tokenisation similarly rely on shared platforms, protocols and service providers that can become systemically important.
- Operational disruptions, cyber attacks or technology failures can therefore have significant implications for the financial system.
Amplification and propagation of stress
Third, AI and digital finance can influence how stress propagates across the financial system:
- The widespread use of similar AI models, data or decision rules can lead institutions to respond to shocks in similar ways, increasing correlations in behaviour. Tokenisation platforms can create new and sometimes complex interdependencies across markets and activities.
- These features can amplify shocks through contagion and procyclicality, allowing stress to spread more rapidly and broadly across institutions, markets and jurisdictions.
Complications for central banks in identifying and managing financial stability risks
These risk channels are not fundamentally new. However, technological change associated with AI and digital finance may increase the intensity, speed and complexity of the flow of these financial stability risks, complicating how central banks and regulators identify, assess and manage them.
First, intensity. As the adoption of AI and tokenisation becomes more widespread, exposures and interdependencies may become more concentrated. Reliance on shared platforms, common data sources, similar models or key service providers can increase the scale of the potential impact when disruptions occur. As a result, shocks that might previously have affected individual institutions could have more pronounced system-wide consequences, raising the intensity of financial stability risks.
Second, speed. Both technologies can accelerate how risks materialise. AI enables faster and more automated decision-making, while tokenisation can shorten transaction chains and settlement processes. Under stress, this may compress the time available for institutions and authorities to respond, as liquidity pressures, operational disruptions or market reactions unfold more rapidly. Faster transmission can therefore make risks harder to contain once they emerge.
Third, complexity. AI and digital finance can also increase the complexity of the financial system, making risks more difficult to observe and assess. In the case of AI, opaque models, unstructured data and reliance on third-party service providers can complicate risk assessment and validation. For digital finance and tokenisation, programmability, composability and the involvement of multiple intermediaries across jurisdictions can create intricate and sometimes opaque interdependencies.
International cooperation, governance and the role of the BIS
The developments I have discussed have a strong cross-border dimension. AI services, digital platforms and tokenisation arrangements often operate across jurisdictions, while responsibility for financial stability remains largely national. This creates a natural role for international cooperation, particularly in the area of governance.
As finance becomes more digital and interconnected, governance frameworks – covering accountability, risk management and oversight – become increasingly important. In the case of AI, governance issues arise around model risk management, data governance and reliance on third-party service providers. For digital finance and tokenisation, governance and design choices – such as access arrangements, operational responsibilities and settlement processes – can have important implications for resilience, especially when activities span borders.
Because these developments cut across jurisdictions, fragmented or inconsistent approaches to governance can create gaps and frictions. This highlights the value of greater alignment and coherence in governance and regulatory frameworks, while respecting national mandates and differences in market structure.
In this context, the Bank for International Settlements plays a supporting role. The BIS provides a forum for central banks to exchange views on governance challenges, supports analytical work to clarify the financial stability implications of AI and digital finance, and facilitates collaboration and shared learning, including through the BIS Innovation Hub. By working with the FSB, International Monetary Fund and other strategic partners, we at the BIS stand ready to help foster sound and interoperable governance approaches that support innovation while safeguarding financial stability.
Conclusion
Let me conclude.
Artificial intelligence and digital finance, including tokenisation, are reshaping how financial markets function. They offer important opportunities to improve efficiency, integration and innovation. At the same time, as these technologies grow, they may intensify familiar financial stability risks – affecting liquidity, operational resilience, interconnectedness and procyclicality.
For central banks, the challenge is not to resist innovation, but to understand how it changes the nature and transmission of risks, and to ensure that governance frameworks and policy approaches remain fit for purpose. By strengthening international cooperation and drawing on shared analysis, central banks can help ensure that technological innovation supports a stable and resilient financial system.
Thank you.
