How AI can help solve the pressing problems of banking and finance

The UNSW FinTech AI Innovation Consortium is a groundbreaking industry-led research group spearheaded by leading AI academics at UNSW Sydney

Credit risk as well as environmental measurement and reporting are areas of significant concern to financial institutions, and artificial intelligence (AI) can play a major role in improving efficiencies and outcomes in these areas from a finance technology (FinTech) perspective.

“These are areas of great concern, to which financial institutions are dedicating a lot of resources,” said Fethi Rabhi, a Professor in the School of Computer Science and Engineering at UNSW Sydney. Credit risk and environmental measurement and reporting are highly data-driven, technically intensive and regulated, and Prof. Rabhi said they have historically relied heavily on human judgement.

This has introduced the potential for bias or compromised data quality, which can affect AI outcomes. “There is currently a lack of understanding of exactly what these implications are going to be. The only way to solve it is for people with experience to keep hammering away on it together,” said Prof. Rabhi, who was speaking following the recent launch of the UNSW FinTech AI Innovation Consortium – an industry-led research group spearheaded by leading AI academics at UNSW Sydney’s Faculties of Engineering and Business together with founding industry partners Westpac, Amazon Web Services, Databricks, BrewAI and Cognitivo Consulting.

He observed there is a pressing need from industry for ready-made AI solutions that are adapted to their needs. “Such solutions must bring value to the organisation and fit within the organisation’s IT strategy and their established processes concerned with software development, maintenance, data governance and evolution,” said Prof. Rabhi, who added these solutions must leverage existing IT and data assets while complying with relevant regulations and satisfying ethical considerations.

For example, replacing a trading decision-support system with an AI box may deliver monetary value, but he said this introduces the issue of how to explain its internal working to regulators as this is an essential requirement before deploying such a system. “Providing ‘explainability’ in AI systems is a complex problem as it needs to take into account not only the method used but also the application domain in which it evolves,” he said.

A multidisciplinary approach is needed to tackle AI challenges 

Prof. Rabhi, who has spent more than two decades involved in industry research projects bringing together software engineering and finance, observed that research in the AI field is very fragmented, with different researchers looking at AI applications from many different angles.  

For example, he said researchers from the information systems field are looking at how adoption issues or how companies can implement effective AI strategies. Researchers from the computer science field, however, are looking at efficient machine learning methods to crack specific problems or data access and management issues. 

“In many cases, researchers are making simplifications or assumptions depending on the field of study,” he said. “For example, if the focus is on machine learning methods, researchers will use clean data that has already been pre-processed. If the focus is on strategy, the details related to technology choices will not be addressed.” 

This means that building effective AI solutions in a specific context can only be accomplished via a multidisciplinary effort where different experts contribute their expertise from different areas. “This has to be accompanied by a rethink of the traditional methods we currently use to build software as they need to be adapted to become a collaborative effort between multidisciplinary domain experts and software development/data management teams.” 

Fintech AI innovation at Westpac

David Walker, group chief technology officer at Westpac (which made an initial investment of $230,000 over three years to fund PhD students as part of the consortium) said the work of the alliance would be ground-breaking in helping to accelerate and scale the next generation of AI for the finance sector. 

“There are a lot of AI capabilities that we’re really holding back until we’re confident we can let it go,” Mr Walker said during a session at TechX, Westpac’s internal five-day technology convention. “Currently all of the artificial intelligence applications across Westpac are ‘supervised’, but ‘unsupervised’ AI is advancing at pace. We need to understand how we put the guardrails, the controls in place to make sure that any AI is acting in a responsible, ethical way, that it’s fair and that there’s no bias built into it.

“We have a real journey ahead of us to get there, but we’ll be working with some of the leading thinkers in this country and globally through this consortium to help us break this down and solve it,” said Mr Walker, who added Westpac was keen to contribute to and benefit from rapid developments in AI.

“AI is becoming more and more advanced and with that is a need to ensure it is used in the right way, ‘AI for good’. We are incredibly excited to participate with UNSW and tap into the expertise of the consortium to solve important and complex problems across the sector, from helping customers to reducing risk and supporting businesses with their sustainability priorities," he said.

How to solve uniquely complex data problems with AI 

The consortium is part of the UNSW UNOVA Research Lab and sits under the umbrella of UNSW’s AI Institute, which aims to boost the impact of around 300 of UNSW’s AI, machine learning and data science academics. “It’s almost like a one-stop shop for the industry – for businesses who know they want to use AI and have problems to solve, and we will work to solve those problems,” said Prof. Rabhi, who is a lead for the consortium together with Felix Tan, Associate Dean International and UNSW UNOVA founder, and Eric Lim, Head of Fintech Research at UNSW UNOVA and Senior Lecturer in the School of Information Systems and Technology Management at UNSW Business School.

Dr Lim added that the combination of AI with FinTech is an exciting amalgamation as it represents an extension of physical lives into the virtual domain. “There have been extensive discussions about the ethics associated with AI for a long time," he said. 

"For AI to serve humanity, FinTech could serve as humanity’s extension to AI in providing actual stakes from which AI could learn and be actively directed by the mentorship of humans in making decisions that are consequential and impactful. Such a paradigm has a positive self-reinforcing effect that will improve the technological design of both fintech and AI that will serve as critical infrastructure for the new Web3 economy.”

PhD students specialising in AI have been recruited to tackle the problems brought by the consortium industry partners, and their aim will be to collaborate intensively to advance the state of AI knowledge and share their findings with the broader AI community. In the process, Prof. Rabhi said the consortium has developed several market innovations in the way software development practices have evolved. These include collaborative tools or platforms, new ways of working for multidisciplinary teams, leveraging new cloud-based infrastructures, new ways of modeling knowledge for the purpose of sharing, and new ways of acquiring and managing large amounts of data. 

“This is why the consortium includes not only industry partners that have interesting problems to solve but also technology partners that offer innovative platforms and tools that can be included in the overall solution,” he said. “That’s what the consortium does – we’ll hammer away at very specific problems intensively, and each project will pave the way for the next, and that’s the only way you can make an impact.” 

Subscribe to BusinessThink for the latest research, analysis and insights from UNSW Business School

The consortium welcomes new industry partners and academics (from within or outside UNSW) to contribute to existing projects or propose new projects that fit its vision and mission statement. The consortium also welcomes the exchange of ideas with other related research projects and initiatives. If interested in engaging with the consortium, contact FAIC@unsw.edu.au.