(Yicai) June 23 -- At a time of shifting economic dynamics and increasing geopolitical complexity, the Lujiazui Forum provides a critical platform for shaping the future of global finance. This year’s theme - Advancing Financial Development and Cooperation under Global Governance Initiatives: New Vision, New Challenges, and New Opportunities - captures both the scale of ambition and the reality of the moment. Financial systems are becoming more interconnected, more digital, and more dependent on the effective coordination of policy, infrastructure, and innovation across borders.

Among the forces driving this transformation, artificial intelligence stands out. AI is rapidly moving from experimentation to implementation, with financial institutions exploring how it can improve efficiency, enhance insight, and support better decision-making. Its potential to accelerate financial development - whether through more sophisticated risk analysis, improved market transparency, or broader access to financial services - is increasingly clear.

Yet the opportunity presented by AI is inseparable from a deeper structural challenge - one that will ultimately determine whether it delivers on its potential: can it be trusted?

In a global financial system that depends on confidence - between market participants, across jurisdictions, and between institutions and regulators - AI introduces both new capabilities and new questions. As financial services become more reliant on automated systems, the quality, transparency, and governance of the underlying data becomes more important than ever. Without this foundation, the promise of AI risks being undermined by concerns over reliability, accountability, and systemic risk.

This is why the sustainable development of AI is not simply about technological advancement. It is about the conditions required to enable that advancement safely and sustainably. In this context, trusted data emerges as a central pillar - linking innovation to governance, and opportunity to responsibility.

AI’s promise - and its dependency

AI’s capabilities are advancing rapidly. Large language models (LLMs), machine learning algorithms, and AI agents are already transforming how financial professionals work - simplifying workflows, automating complex tasks, and enabling faster access to information.

These tools are redefining productivity: enabling natural language interactions with data, accelerating analytics, and delivering insights in real time. The implications are profound. AI has the potential to improve market efficiency, enhance risk management, and broaden access to financial services.

But AI systems are only as good as the data that underpins them.

Data is the raw material of AI. It determines how models are trained, how they perform, and how reliable their outputs are. In consumer applications, imperfections in AI outputs may be tolerable. In finance, they are not.

Financial professionals operate in an environment where precision is critical, accountability is essential, and errors can carry significant consequences. Decisions about capital allocation, risk exposure, or regulatory compliance cannot be based on approximations or incomplete information. As a result, the effectiveness of AI in financial services is inseparable from the quality, provenance, and integrity of the data it uses.

Beyond volume: the case for “trusted data”

Much of the recent discourse around AI has focused on scale - the vast quantities of data required to train increasingly sophisticated models. While scale matters, it is not sufficient.

Trusted data goes beyond simply being large in volume. It is defined by several critical characteristics:

· Accuracy – data must be precise and reflect reality, particularly in time-sensitive markets

· Completeness – datasets must capture the full context required for decision-making

· Provenance – the origin of data must be clear, enabling users to understand where it comes from and how it can be used

· Timeliness – information must be current and relevant

· Governance and compliance – data must be managed in line with regulatory requirements and usage rights

These attributes are essential in a sector where transparency and auditability are not optional. Regulators, institutions, and end-users all require clarity. They need to understand not just what an AI system has concluded, but why.

While publicly available data can be useful, it is often insufficient to meet these standards. Many AI models trained on open internet data can generate plausible outputs, but these outputs may lack the reliability required for financial decision-making.

Financial institutions rely heavily on proprietary, historical, and permissioned datasets - often combined with rich metadata - to ensure completeness and traceability. Without these elements, AI outputs risk becoming less dependable, even if they appear sophisticated.

The concept of “trusted data” is also increasingly reflected in China’s policy direction. Initiatives to build trusted data spaces aim to facilitate the secure sharing and utilisation of data across institutions, while maintaining clear rules on data ownership, usage rights, and compliance. As these frameworks evolve, they are expected to support more efficient data flows and unlock greater value from data assets, particularly in highly regulated sectors such as finance.

The Chinese government has placed increasing emphasis on the governance and strategic value of data as a production factor. Efforts to develop “trusted data spaces” and strengthen data infrastructure reflect a broader policy focus on enabling secure, compliant and efficient data circulation. These developments are particularly relevant for financial markets, where confidence in data quality and governance underpins both domestic market development and cross-border participation.

Well-known as a global financial markets infrastructure, the London Stock Exchange Group (LSEG) is also a leading data provider. We pride ourselves on data quality and the depth and breadth of our data is unmatched. We offer the financial sector’s largest portfolio of real-time, pricing, reference, time series, and machine-readable content, as well as company, research, news and ESG data. Not only do we license data from hundreds of varied sources – both public and private – and work with expert partners worldwide, but more than half of our data is proprietary. We legally own the data content, the models and the methodologies, and we have a proprietary taxonomy that is embedded in customer applications.

All our datasets are enriched. From proprietary symbology and identifiers to standardisation and normalisation, tagging, sentiment analysis, point-in-time and analytics, we create added value for our customers throughout their workflows. Our real-time data is deeply integrated in our customers’ workflows and delivered over a dedicated network, and we provide unique datasets which are not replicable, like Tick History – going back 30 years across more than 90 million securities.

LSEG offers customers far more decision-useful data, and far more value, than merely accessing data from publicly available sources.

Trust as a systemic issue

The importance of trusted data is not confined to individual firms or use cases - it is a systemic issue for the global financial ecosystem.

As AI becomes more deeply embedded in financial markets, its outputs will influence everything from pricing and liquidity to risk models and capital flows. This creates a new layer of interdependence across institutions and markets.

If the data feeding these systems is inconsistent, unreliable, or fragmented, it has the potential to amplify risks. Conversely, if the data ecosystem is robust, interoperable, and well-governed, AI can enhance stability and efficiency.

This is where the theme of global cooperation becomes particularly relevant.

Ensuring the availability and integrity of trusted data requires coordination across jurisdictions. Data flows must be both secure and sufficiently open to support innovation. Regulatory approaches must be aligned enough to enable interoperability, while still reflecting local priorities.

In this context, China’s efforts to strengthen data infrastructure and governance frameworks are also contributing to the broader global conversation on interoperability. As financial markets become increasingly interconnected, aligning approaches to data standards, transparency, and governance will be key to enabling trusted cross-border data flows and supporting global financial stability.

Striking this balance is not straightforward. But it is essential if AI is to support a more connected and resilient global financial system.

Governance: enabling innovation while managing risk

The rapid pace of AI development has prompted policymakers and regulators to act. Around the world, frameworks are emerging to guide the responsible use of AI - addressing issues such as safety, accountability, fairness, and privacy.

In China, this growing reliance on high-quality data is also being matched by a more systematic approach to data governance. Policymakers have been advancing frameworks to support data classification, security, and cross-border data flows, with the aim of balancing innovation and risk management. This evolving regulatory landscape is an important step in building trust. But regulation alone is not sufficient.

Businesses also have a critical role to play. They must develop robust governance frameworks that ensure AI systems are used responsibly, transparently, and in line with both regulatory expectations and societal values.

Effective AI governance typically rests on several key pillars:

· Clear accountability structures, ensuring responsibility for AI outputs is well-defined

· Robust risk management processes, identifying and mitigating potential harms

· Transparency and disclosure, making it clear when AI is being used and how outputs are generated

· Bias monitoring and mitigation, to ensure fair and equitable outcomes

· Data protection and privacy safeguards, protecting sensitive information

These principles are becoming widely recognised across the industry, reflecting a growing consensus on what constitutes responsible AI.

However, organisations need time to operationalise them - embedding governance into workflows, adapting to different regulatory regimes, and ensuring consistency across jurisdictions.

In this context, continued engagement between policymakers and industry will be essential. Regulation should enable innovation while maintaining trust - not stifle it.

Unlocking opportunity through trusted data

If the challenges are significant, so too are the opportunities.

When underpinned by trusted data, AI has the potential to deliver meaningful benefits across financial services – be it improved decision-making, through deeper insights and more comprehensive analysis; greater efficiency by reducing manual processes and operational complexity; and enhanced risk management.

These benefits extend beyond individual institutions. They have the potential to support economic development, improve market resilience, and enhance global financial integration.

But these outcomes are not guaranteed.

Realising them will require sustained investment in data quality and infrastructure. It will require collaboration across public and private sectors. And it will require a shared commitment to trust as a guiding principle.

A collective responsibility

Ultimately, building trust in AI is a shared responsibility. Policymakers must provide clear, consistent frameworks that balance innovation with protection. Businesses must demonstrate accountability, transparency, and responsible use. Data providers and technology firms must ensure that the information underpinning AI systems is reliable, well-governed, and fit for purpose. And we must all help coordinate efforts and promote interoperability across borders.

This collective approach will be critical in navigating the “new vision, new challenges, and new opportunities” that define the current moment.

Conclusion: trust as the enabler of scale

AI is often described as a revolutionary technology - and rightly so. But its success in financial services will not be determined by technical capability alone. It will be determined by whether it earns the trust of those who use it.

Trusted data sits at the heart of that equation. With it, AI can deliver on its promise - transforming financial services in ways that are both innovative and responsible.

As the global community looks to advance financial development and cooperation, investing in trusted data is not simply a technical requirement. It is a strategic imperative.

Because in finance, as in AI, trust is everything.

The author is Nicole Chen, Market Head of North APAC, Country Head of China, London Stock Exchange Group (LSEG).