Fintech is no stranger to artificial intelligence (AI).
According to estimates, the market for AI in fintech will reach $31.71 billion in 2027, growing at a rate of 28.6%. Already, AI is being applied in a number of different use cases. According to Cambridge Centre for Alternative Finance, 90% of fintech companies already use AI.
But a new beast has risen, generative AI, which promises to revolutionize how we approach financial services.
However, right now, the industry is merely scratching the surface, and there is a long way to go before generative AI can meet its full potential.
“How is Gen AI being used by financial services?” said Robert Antoniades, Co-Founder and General Partner of Information Venture Partners. “The simple answer is it’s not being used. Certainly not broadly. But what Gen AI has done is it has increased the recognition of the power of AI for financial institutions.”
He explained that while companies have implemented Generative AI tools like Chat GPT to help streamline customer-facing processes, the technology could greatly impact the back end of financial services.
But for this to be possible, the technology would have to be 100% accurate. For the time being, it certainly is not.
The Road to Accuracy
An example of how inaccurate Gen AI can be was glaringly obvious just a week ago.
On June 1, social media was a flurry of excitement as “anonymous sources” had apparently told journalists that SEC Chairman Gary Gensler had stepped down pending “an internal investigation.” Hours later, these claims were debunked.
So who was the journalist who made these false claims? A generative AI bot.
“You have to understand that in financial services if it’s anything important, it has to be 100% accurate,” said Antoniades. “There’s no room for hallucinations. There’s no room for errors. AI-generated answers are fascinating to see because they’re actually decent, but they are not accurate.”
“For the purposes of prospecting or marketing, it’s fine. But for financial advice, no, absolutely not. For record keeping. Absolutely not.”
Inaccuracy could have devastating consequences.
The need for perfection becomes clear when considering potential use cases for Generative AI in financial services.
Financial advice has been an area many have singled out as a focus for generative AI in finance. Due to cost, financial advisory services are an area that is inaccessible to many. However, the use of Generative AI could change that, tailoring advisory services to specific customer requirements based on their interaction.
“Gen AI is actually a very interesting use case there of how to provide that interaction and contextualization between the customer and the financial institution,” said Antoniades. “By ingesting all that data, it can now have what one would consider a conversation with a client.”
“If you think about what you do today, you call into a call center. You look on their website. You listen to friends, you do web searches, and you can actually have an interaction. And I think the beauty of generative AI is the generative part. It’s the ability to actually converse.”
In addition, fraud and AML have been singled out as areas that could be significantly improved and already have increasingly applied AI and machine learning models to improve outcomes.
However, Antoniades explained that a particularly disruptive application of the technology could be in modernizing infrastructure.
The banking system is built on infrastructure that has remained the same for years. Powerd by COBOL, a language developed in 1959, it has stood firm while new tech popped up and sped past.
The creaking framework is, however, showing its age. New programmers have turned away from learning the language favoring more universal code, and the structure requires extensive custom programming in order to make changes.
“I think about this as a way to modernize infrastructure,” said Antoniades. He explained that Generative AI could be used to write the outdated COBOL code as well as provide a kind of patch to accelerate a shift to new infrastructure.
“For a financial institution to make the transition off to a new platform is a risk. But I think Gen AI can now replace old infrastructure, old code with new code and help a financial institution transition to the modern era.”
The consequences of an error in these legacy frameworks could, however, be fatal.
“When you make a deposit in your bank account, you want to know the money’s there,” said Antoniades. “It’s not that it can be there 99.9% of the time. It’s always there. When they give you advice, they really should be 100% accurate. It shouldn’t be 90% accurate.”
The potential is there, and Antoniades said financial institutions are aware of it. All the industry needs now is enough development for GenAI’s outcomes to be close to perfect.
RELATED: AML efficiency from networks and AI