Artificial intelligence has made it possible to extract critical data from unstructured sources at speed and at scale. But the headlong rush to adopt the sorts of tools that can mine this rich vein of information is exposing organisations to new risks.
Generative AI, whose models are commonly applied to trawling PDFs, emails, financial reports and other documents that previously would have been pored over by team members, is prone to well-documented biases and faults.
If the data that’s extracted is contaminated for any reason, the decisions, workflows and products that use it will be similarly compromised.
At this year’s A-Team Group Data Management Summit London, high-level speakers from across the data landscape will discuss this and a host of other issues facing data chiefs and engineers within the institutional space.
Unstructured Data Extraction Presents New Risks
Joanne Biggadike, head of data governance at Schroders, will moderate a panel discussion that will focus on the risks that organisations face as they adopt or investigate AI applications. The gathering will also look at some of the solutions that can help them avoid AI-related risks.
Biggadike told Data Management Insight that many companies are driven to AI adoption under a pressing sense of need.
“Many organisations are feeling the pressure — whether from a desire to be leaders in new technologies or simply to demonstrate to their customers and clients that they are proactively adopting AI,” Biggadike told Data Management Insight. “Terms such as ‘being left behind’ or “out of date” often do little to alleviate this pressure. In highly regulated financial industries, however, the drive for innovation must be closely balanced with the need to act safely and responsibly.“By its very nature, venturing into new ways of working brings an element of risk and can challenge existing notions of trust.”
The session will also comprise Paul Barker, chief control office, enterprise technology at HSBC Group; Michael Dimopoulos, driving global markets, client reporting solutions and digital innovation at BNP Paribas; and, Nicole Hansen, compliance and conduct change lead at NatWest Group.
“In this panel, we will be discussing how to strike the right balance between innovation and governance, and how to ensure that progress does not create more problems than it solves,” said Biggadike.
“This is especially important when considering the use of GenAI: does the use case truly make sense, is it financially viable and does your organisation have the technical capability to implement it in line with your approved ways of working?”
Data Products and the Strengthening of System Capabilities
Many institutions are turning to data products made available on internal and external marketplaces to provide new capabilities to their data management architectures. These relatively new phenomena are offering low-code/no-code, self-service opportunities for firms to create and then deploy new functionality that they may not be able to find offered by vendors.
Again, however, their use has presented organisations with potential new risks, both technological and operational. Are the products being properly designed for existing business needs? How can they be governed to ensure they comply with corporate and regulatory guidelines? How does an organisation build a data product framework at scale?
These are some of the questions that a separate panel will digest. Moderated by Niresh Rajah, an expert chief data and AI officer as well as board adviser/NED, the panel will also feature Mridula Mutharaju, head of data and analytics, commercial and institutional at NatWest Bank; Effie Kilmer, chief data and AI adviser at Microsoft EMEA; Peter Jackson, chief data officer at Schroders; and, Jon Asprey, field chief data officer at Atacama.Adoption Requires a Joined-Up Approach
NatWest’s Mutharaju told Data Management Insight that organisations must take a holistic approach to the use of data products.
“We need to start with business strategy and customers first, and then understand the value that can be unlocked through the use of data products that provide modelled, ‘single version of truth’ data with known quality and lineage,” she said.
“Data products need to be explicitly aligned with business value unlock – outcomes delivered from adoption. Both of these require business, data, and technology teams to work closely together, so it’s an enterprise activity that just happens to be about data.”
- A-Team Group Data Management Summit London will be held at etc.venues, 8 Fenchurch Place, London on March 26. To book your place, click here for registration details.
Subscribe to our newsletter
