Skip to content

Asset Tokenisation, Artificial Intelligence, Quantum Computing – PZF 2024

Ravi Menon

Ambassador for Climate Action, Singapore

 

Keynote Address at the Point Zero Forum

Zurich, Switzerland

2 July 2024

 

Ladies and gentlemen, good morning. It is a pleasure to be back in Zurich for the Point Zero Forum. I look forward to the conversations over the next two days.

 

Let our goal be nothing short of audacious and bold: to apply the best of technology to re-shape the financial ecosystem of the future: more efficient, more value-adding, more inclusive.

 

I think three technologies have the best prospects for transforming finance:

  • asset tokenisation
  • artificial intelligence
  • quantum computing

It is not a surprise that much of the PZF programme is devoted to these three topics.

 

Asset Tokenisation

 

Let me start with tokenisation. Together with distributed ledgers, it has the potential to create a more efficient global financial architecture.   By this, I mean a global network of interoperable systems that allows payment, clearing, and settlement to take place instantaneously and seamlessly.

 

Tokenisation is essentially about representing the ownership rights over any asset as a digital token. It has two critical features that can fundamentally transform the nature of financial transactions: direct exchange and fractionalisation.

  • Direct exchange without the need for intermediaries makes possible atomic settlement or the simultaneous exchange of two assets in real-time. It eliminates settlement risk, duplicative reconciliation, and the need for large funding accounts.
  • Fractionalisation of assets makes possible the partial collateralisation of assets and grants a much broader range of investors access to these assets.

 

Tokenised financial assets are proliferating rapidly.

  • We are tokenising foreign exchange for a 24/7 global liquidity pool.
  • The Bank of New York Mellon and OCBC Bank are trialling a cross-border foreign exchange tokenised solution to enable secure payments across different networks.
  • We are tokenising bonds for seamless cross-border distribution and settlement.
  • UBS, SBI Digital Asset Holdings, and DBS Bank executed a pilot repo agreement with natively issued digital bonds, working across Switzerland, Japan and Singapore
  • We are tokenising funds for efficient issuance and trading.
  • JP Morgan is converting money market shares into digital tokens, which then serve as collaterals for over-the-counter derivative deals.

 

Policy impetus to facilitate digital assets is gaining traction.

  • The Monetary Authority of Singapore, along with partners from the financial industry and the International Monetary Fund (IMF) as observers, has launched an initiative called Project Guardian to lay down frameworks and protocols to harness the benefits of tokenised assets.
  • Project Guardian aims to integrate digital assets seamlessly into mainstream financial operations, like listing, distribution, trading, settlement and asset servicing.

 

It is not only assets that can be tokenised, money itself can be tokenised.

  • In fact, money needs to be tokenised - to enable the payment, clearing, and settlement of tokenised assets simultaneously on the same platform.  
  • But like all other money, tokenised money must have stability of value and security of redemption.

 

There are four contenders for tokenised money.

  • central bank digital currencies or CBDCs
  • tokenised bank liabilities
  • well-regulated stablecoins
  • private cryptocurrencies

Private cryptocurrencies have failed the test of money. They have performed poorly as a medium of exchange or store of value beyond that for purposes of speculation.

 

CBDCs, tokenised bank liabilities, and well-regulated stablecoins are much more promising as digital money.

  • SWIFT has launched the testing phase of a CBDC interoperability connector that links the CBDCs of three central banks — the Hong Kong Monetary Authority, the National Bank of Kazakhstan, and the Reserve Bank of Australia — and 30 financial institutions.
  • The aim is to enable tokenised currencies to co-exist with fiat-based currencies and payment systems.

 

Whether tokenisation achieves its potential to transform finance depends on the underlying networks on which digital token operate. These digital asset networks are in the form of distributed ledgers which enable the ownership and transfers of tokenised assets and tokenised money to be recorded consistently among participating entities and directly on the ledger.

 

But existing digital asset networks are not fit-for-purpose as a global infrastructure for a tokenised financial system.

  • Currently, most digital asset networks are not interoperable, reflecting differences in commercial motivations or legal and regulatory requirements.
  • Public permissionless blockchains have attracted many users and applications but they suffer from lack of accountability, anonymity of service providers, and legal uncertainty.
  • Private permissioned blockchains address regulatory and legal concerns but suffer from lack of interoperability, leading to fragmentation of liquidity in digital assets.

 

To realise the vision of seamless financial transactions globally, we need open and interoperable digital asset networks that are compliant with regulatory requirements. We cannot force consolidation of all financial transactions onto a single network. It is more feasible to work towards making these diverse networks interoperable.

 

This is easier said than done. But there are a couple of promising efforts internationally.

 

The Bank of International Settlement (BIS), for instance, has proposed the concept of the "Finternet" as a vision for the future financial system: multiple financial ecosystems interconnected with one another through unified ledgers – much like the Internet.

 

The Monetary Authority of Singapore, together with industry partners like BNY Mellon, Citi, Société Générale-Forge (SG Forge), JP Morgan and MUFG, has initiated the Global Layer One or GL1 project.

  • This is a global public good digital infrastructure, through which cross-border transactions and trading of a range of tokenised assets can be done seamlessly.
  • whitepaper which details the design principles, objectives, considerations and potential uses of GL1 was published just last week.

 

The financial architecture of the future may well be one where a tokenised financial system co-exists seamlessly with a more traditional one. It is a vision that is by no means certain but well worth realising – for what it means for economic efficiency, expanded opportunity, and financial inclusion.

 

Artificial Intelligence

 

The second megatrend in technology that will transform finance is artificial intelligence or AI. It encompasses a diverse range of technological models aimed at replicating human intelligence in various tasks.

 

Deep Learning Models analyse multiple layers of complex data to train artificial neural networks which can comprise millions of densely interconnected processing nodes.

  • Ant International uses deep learning to assess the creditworthiness of loan applicants by analysing thousands of data points about their online behaviour and digital footprint.

 

 

Generative AI or Gen AI uses various AI technologies to create new content, including

text, images, and music. It is being increasingly deployed in the financial industry.

  • Take for instance fraud detection. Gen AI can track transactions based on location, device, and operating system, flagging any anomalies or behaviour that does not fit expected patterns.
  • Another example: personalised financial advice. Gen AI can make customised budgeting, saving, and investing recommendations based on customers' financial goals, risk profiles, income levels, and spending habits.

 

AI has great potential to enhance financial services. But to realise this potential, we must get not just the technology right but also the governance right.

 

There are four key governance issues in AI:

  • One, privacy of data. How do we harness the benefits of data aggregation and data sharing while safeguarding confidentiality of personal data?
  • Two, explainability of results. How do we minimise the black box syndrome? The massive amount of data, complexity of the algorithms, and dynamic nature of AI systems make their results difficult to interpret and explain. Very much like a human brain.
  • Three, accountability for decisions. How can we hold humans ultimately responsible for decisions made by self-learning algorithms and machines? You cannot charge a machine in a court of law. You need to charge a person.
  • Four, acceptability of outcomes. How do we minimise unconscious bias, social exclusion, and ethically unacceptable outcomes? AI models trained on incomplete or biased data can generate seemingly plausible but incorrect content or unsound predictions. These can in turn lead to flawed financial decisions regarding credit or investments.

 

We need regulation of AI; we need it fast and we need it harmonised. The technology is developing rapidly and spreading widely. More than 37 countries have proposed AI-related legal frameworks. The approaches taken by the EU, the US, and China offer interesting contrasts, and things to learn from.

 

The EU has distinguished itself as a global leader in AI regulation.

  • The EU Artificial Intelligence Act was passed by the European Parliament in March 2024. It represents the first comprehensive regulation on AI by a regulator, anywhere.
  • The Act establishes binding rules for general-purpose AI models and institutes a centralised governance structure at the EU level through a dedicated European AI office.
  • The EU AI Act is based on a rights-driven approach.
  • It champions a risk-based approach, imposing regulatory burdens only when an AI system is likely to pose high risks to fundamental rights and safety.
  • The Act is human centric, mandating a "human-in-the-loop" approach that requires AI designers to engineer systems which enable human intervention or control when necessary.
  • The Act mandates transparency in AI systems, which means users have the right to understand how these systems reach decisions.

 

The United States does not yet have comprehensive AI regulation.

  • Over 400 new bills have been debated in state legislatures this year. The majority of these are aimed at regulating smaller slices of AI, for example, rules around deepfakes.
  • At the federal level, AI governance in the US primarily comprises of voluntary commitments and non-binding measures, underpinned by the principle of responsible innovation.
  • In July 2023, the Biden Administration convened 7 AI companies to the White House - Amazon, Anthropic, Google, Inflection, Meta, Microsoft, and OpenAI. The Administration announced that it had received voluntary commitments from these companies on ensuring that key principles of safety, security and trust were met when developing AI.

 

China has taken a targeted and phased approach towards AI regulation.

  • It started in 2017 with a high-level roadmap on the country’s approach to developing AI technology and applications - with goals up to 2030.
  • China has since produced several interesting technology-specific AI regulations.
  • It has regulations on algorithms, to prohibit excessive price discrimination and to protect workers subject to algorithmic scheduling that imposed unrealistic timelines and workloads.
  • It has rules for deep synthesis, which requires prominent labelling for synthetically generated content.
  • It has rules on Gen AI, which require training data and model outputs to be “true and accurate” – though this may represent a difficult hurdle for AI chatbots.

 

Singapore has focused on providing guidance on the responsible use of AI as well as an AI governance testing framework.

  • The Model AI Governance Framework provides detailed guidance on ethical and governance considerations for deploying AI systems, covering principles such as transparency, explainability, fairness, well-being, and safety.
  • AI Verify provides an AI governance testing framework and software toolkit to help organisations validate their AI systems against 11 AI ethics principles, aligning with international standards and Singapore's own Model AI Governance Framework.

 

Eventually, we will need a robust international framework for the responsible development and use of AI. Such a framework should:

  • disseminate globally information on advances in AI;
  • promote convergence of international AI regulations through dialogue and consensus-building mechanisms;
  • encourage the fair distribution of benefits arising from advancements in AI; and
  • ensure a more secure global environment by mitigating dual-use risks.

 

Quantum Computing

 

The third big trend in FinTech is more nascent: quantum technology. Leveraging the principles of quantum mechanics, this technology has the potential to revolutionise various fields through its radically new approaches.

 

Quantum technology has strong potential applications in financial services. Two key attributes of quantum computing have been particularly useful:

  • One, quantum optimisation. Quantum computers can optimise complex financial processes like portfolio management, risk modelling, trading strategies, and fraud detection. This comes from the technology’s inherent ability to conduct calculations with a much larger set of boundary conditions.
  • Two, quantum simulation. Quantum computers can simulate complex financial systems and models more accurately than classical computers.
  • Investment banking teams can create full digital twins of a bank’s positions that they could use to simulate various macroconditions and pathways.
  • They can perform granular simulations to see how different scenarios would affect every single asset in the bank.
  • This can help optimise capital allocation with regard to collateral and assets.

 

But quantum computing also poses substantial - even existential - risks. The very integrity of financial infrastructure, communications, and data is at stake. Quantum technology can be used to compromise the encryption protocols that safeguard financial data, potentially undermining the fundamental pillars of finance – trust and stability.

 

Let me highlight just a couple of examples of these encryption-related risks posed by quantum computing:

  • one, compromise of inter-bank system interfaces (e.g. open banking APIs) which could allow unauthorised access to sensitive financial data, including personally identifiable information;
  • two, compromise of public key cryptography in wholesale payment systems heavily rely on for authentication which could enable fraudulent payment transactions or data tampering.

 

Embracing quantum technology safely requires a proactive approach to managing its profound security risks.

  • Singapore and Japan are collaborating on the NQSN+ Digital Connectivity Blueprint. This is Southeast Asia’s first quantum-safe network infrastructure that aims to standardise Quantum Key Distribution protocol by enabling two parties to produce a shared random secret key for secure encrypted communication.
  • In the United States, the Quantum Computing Cybersecurity Preparedness Act 2022, mandates the identification of vulnerable federal government information technology, and its migration towards “quantum safe” equivalents.

 

Financial regulators are also becoming increasingly cognisant of quantum computing risks.

  • The BIS Innovation Hub's Eurosystem Centre, with Banque de France and Deutsche Bundesbank, launched Project Leap to enable central banks and the global financial system to transition to quantum resistant encryption.
  • The Monetary Authority of Singapore has issued an advisory to financial institutions recommending they prioritise migrating to quantum-resistant cryptography and collaborate with vendors to work towards making their IT systems quantum-resistant.

Conclusion

 

Let me conclude.

 

The global FinTech wave that began some ten years ago continues to grow. The financial services industry may well be on the cusp of even more transformative change, powered by asset tokenisation, artificial intelligence, and quantum computing.

 

There are deeply impactful benefits to harness from these technologies. But that depends critically on how we manage the substantial risks associated with these technologies.

 

The answer lies in us, not the technologies.

 

I wish everyone fruitful deliberations over the next two days.