The Reuters news staff had no role in the production of this content. It was created by Reuters Plus, the brand marketing studio of Reuters.
Produced by Reuters Plus for
Data integrity and responsible innovation in AI
As artificial intelligence reshapes financial markets, LSEG is leading the way in ensuring that data quality, trust and compliance remain at the forefront of deployment. With its vast datasets and innovative data passport concept, the company is positioned to drive responsible AI adoption while navigating complex regulatory landscapes.
By Satvinder Singh, Group Head, Data & Analytics, LSEG
The power of generative AI (GenAI) has the potential to be transformational for the world’s financial markets, where data is the raw material used for trading trillions of dollars in stocks every day. The capacity of new foundation models to process 75,000 or more words of text per minute has increased expectations of the far-reaching insights that GenAI might deliver. Customers are increasingly demanding more data—especially textual data—at exponentially faster rates. They seek to leverage AI to extract insights from documents much more quickly than humans ever could.
LSEG has long been working with automation and machine learning, but the evolution of GenAI services powered by foundation models has changed what is possible. Today’s models can examine broader planes of information and carry out repeated tasks while reducing instances of human error and biases that affect human judgement. This generates enormous productivity gains and can allow us to see relationships that were not apparent before.
But caution is needed. The growing complexity of contracts and rights management in this field creates inherent challenges in avoiding licensing or contractual breaches, while we know that feeding an AI model with poor-quality information can have calamitous consequences.
While it is often the case that data errors in financial services are unintentional, the impact of the market being misled can be catastrophic. It is not just bankers and hedge funds that are affected; it is all of society.
This is a highly regulated sector for good reason. If you need to fully explain exactly how you made a decision that was based on data, then relying exclusively on a GenAI model is not the best strategy. It simply does not work for all use cases. What’s more, it’s crucial to address the risks of AI hallucinations, model drift and unintended bias. Implementing robust monitoring and validation processes will ensure the reliability of AI-driven decisions.
The key to using AI data is to take the new opportunity to process large unstructured datasets through GenAI models, while ensuring that the data is both trustworthy and licensed. Data in GenAI is largely a quantity game, but it is also a quality game.
LSEG oversees one of the most extensive collections of datasets among data vendors operating in the global financial markets industry. In managing and licensing data for AI, our company leads from the front. We have an open policy that draws from our long experience of working with previous iterations of AI, from automation to machine learning.
Our goal is to streamline the licensing process for our customers. In most areas of our business, we will be licensing the data utilised by customers, rather than the usage itself.
LSEG wants to redefine the future of data in financial services. A 10-year strategic partnership with Microsoft for next-generation data and analytics ensures that we have direct access to Microsoft’s expertise at the forefront of AI research. We also partner with 40,000 customers and have 400,000 end users working with data, feeds, analytics, AI and workflow solutions across 190 markets. Our data estate contains over 100 million instruments from 550 exchanges. Our company data covers 99% of the world’s market cap.
The scale and quality of LSEG’s data—serving a customer base from wealth managers to quantitative investors—is an invaluable security net when working with AI. This access to original and trusted data sources means that, because of the breadth and depth of our data, we do not have to rely on synthetic results, but can instead be confident that our trusted data catalogue will lead us to the right results.
Using our commodities data that goes back to 1990, LSEG customers can ascertain where the market is now and where it has been, and this data provides signals to help model and predict future market moves. LSEG’s data on commodity markets is extensive, reaching from oil and gas to agriculture and weather.
With such quality checks in place, the potential for GenAI to introduce new advances in data-driven decision-making in financial markets is clear.
40,000
400,000
190
LSEG customers
end users
markets covered
instruments in LSEG’s data catalogue
of the world’s market cap in company data
exchanges in portfolio
100m
550
99%
LSEG also operates a comprehensive Tick History dataset of more than 87 trillion ticks, recording stock pricing data from 500 global exchanges over 25 years. But this is not the data that is being targeted by LSEG customers for use in AI models. Rather, it is the company data—the research filings, estimates and other types of documents and data—that provides the company with insights that allow IT to connect the dots.
Feeding this unstructured data into an AI model, however, runs the risk of licence breaches in a rapidly changing global regulatory environment, ranging from GDPR in the European Union to PIPL, which protects personal information rights for Chinese citizens. Licensing and rights of data in financial services has become a complex issue over the last 20 years.
LSEG’s innovative offering is based on the concept of a data passport, so that our customers can see exactly what is contained in a dataset and understand the use cases for which it is licensed. The passport helps customers know exactly what they can and can’t do with the data product they are using.
Founded on the principle of customer transparency, the passport requires a forensic approach from LSEG’s data analysts. We sometimes think of our data products as being like a cake that we are selling, made up of multiple ingredients. We would not want to sell a cake containing nuts to somebody with a nut allergy, and, similarly, we would not want to sell a product containing personal information to a customer without telling them what was in the dataset. Working towards the provision of such metadata is essential to building trust in the integrity of the data itself.
With such quality checks in place, the potential for GenAI to introduce new advances in data-driven decision-making in financial markets is clear. This is a technology that we should be embracing, rather than be scared of. Since one of our biggest challenges as an industry is data discoverability, we need to help our customers find the right data. GenAI can help us with the solutions because it is mind-boggling how much data there is out there.
If this technology is used in combination with trusted, quality data, the impact on financial literacy can be transformational. With the arrival of GenAI, we have an opportunity to democratise this sector. Our ambition must be to ensure that, ultimately, everyone can find the financial services information they need, when they need it, based on their intent.
Learn more about LSEG’s data offerings and how it is managing the future of data here.
The Reuters news staff had no role in the production of this content. It was created by Reuters Plus, the brand marketing studio of Reuters.
Produced by Reuters Plus for
Disclaimer: The Reuters news staff had no role in the production of this content. It was created by Reuters Plus, the brand marketing studio of Reuters. To work with Reuters Plus, contact us here.
Disclaimer: The Reuters news staff had no role in the production of this content. It was created by Reuters Plus, the brand marketing studio of Reuters. To work with Reuters Plus, contact us here.