Informa Australia is part of the Informa Connect Division of Informa PLC
This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 3099067.
21 Apr 2023, by Amy Sarcevic
The recent hype around generative AI has left many business owners and directors in the financial services industry wondering whether they should be using the technology.
While tools like ChatGTP have largely been celebrated in the corporate world, they do raise serious concerns from an ethics and regulatory perspective.
Unlike traditional AI, which finds answers, generative AI invents answers of its own, many of which are unexplainable.
As highlighted in recent headlines surrounding Open AI’s defamation case, some of its answers are also allegedly misinformed.
Theses headlines have sparked particular concern in financial services, where the impact of misinformation can be strong and highly penalised.
Gordon Campbell, Co-Founder of Rich Data Co (RDC) – and a speaker at this year’s Responsible Lending & Borrowing Summit – believes much of this fear is warranted and says it could take years before the financial sector can safely use the technology. The credit domain, in particular, should be weary of its risks, he said.
“Generative AI’s life purpose is – given a specific context – to create something that didn’t exist before, which is equal parts fascinating and alarming when you apply it in the credit domain.
“If the technology is creating something from the context passed to it, the output needs to be validated; but the technology as it stands today is not capable of doing that. So it could not be trusted with keeping borrowers safe,” he said ahead of his speech.
Alongside borrowers, generative AI in its current state would also fail to keep lenders safe, Gordon warned.
“People’s lives can be wrecked in minutes by a bad decision or piece of advice, so pushing out unverifiable information will never be defensible to the regulator. As Microsoft recently positioned, the technology might makes a good Co-Pilot, but it should not (yet) be put in the drivers’ seat.”
While regulation surrounding AI is relatively immature, regulators are sure of some basic principles.
“Two of the things that regulators currently hinge on are explainability and contestibility; and generative AI is currently lacking on both of these fronts.
“In contrast, credit behavioural scoring as a technique was initially developed in 1956 and is still being used today, because you can hold it up in front of anyone and explain how the technology arrived at the answers it did. This is possible today with other types of AI techniques, but not generative AI.”
Regulator side, today’s generative AI would also scare off most risk professionals in the financial services sector, Gordon said.
“I like to draw on that old adage, ‘lawyers love law, but risk people hate risk’. In this moment Generative AI is incredibly dangerous to banks, especially if it being used to help them make money, so I doubt risk professionals will embrace it quickly.”
Despite these concerns, Gordon is confident the technology will eventually earn its place in the finance sector, particularly as the credit gap continues to grow and piques the industry’s appetite for more inclusive access to credit.
“The world is moving at pace – new things are happening week by week.
It’s not like any other time in history. This means, at some stage, the research will point to when and where generative AI can be used within financial services; and some guardrails will be developed. People will eventually invent safe uses – but right now is not that time,” he said.
In the broader financial services ecosystem, Gordon says the technology is already being used quite safely and will continue to flourish as more use cases emerge.
“Most software houses will already be using ChatGPT-style technology to develop core code or operational scripts. The technology is really good at finding examples of codes, snippets, or even entire programs and adapting them. It should not write the program, but could be used to kick-start it,” he said.
“It is also good at helping people find simple words to describe complex concepts, but, in my opinion, should never be used for final copy. Content needs human input in order to not be bland.”
Sharing more insights about the use of generative AI in financial services, Gordon Campbell is due to present at the Responsible Lending & Borrowing Summit on 22-23 May at the Radisson Blu Plaza Hotel, Sydney. Learn more and register your place here.
About Gordon Campbell
Gordon is a co-founder at Rich Data Co, an Australian FinTech using AI to help banks make better and more inclusive credit lending decisions. He has grown the company over the previous seven years to provide an innovative AI Decisioning in the credit domain to major banks. He has previously been involved in product strategy and product development within three international software organisations, driving new capabilities from inception to market.
Informa Connect Australia is the nation's leading event organiser. Our events comprise of large scale exhibitions, industry conferences and highly specialised corporate training.