Hey ChatGPT. What’s the buzz around AI?

Artificial Intelligence is certainly not new to the finance industry, with credit card fraud detection being an early use of the technology.

But the November launch — and subsequent worldwide buzz — over OpenAI’s ChatGPT and its $10 billion investment from Microsoft is firing up the financial services industry as stakeholders see the technology’s potential.

“ChatGPT has merely highlighted that we need to think even bigger about: how AI can fit into the financial industry,” Mike Lefebvre, director of cybersecurity at fintech SEI Sphere, told Bank Automation News.

“ChatGPT effectively became a household name within one workweek and has made waves in almost every industry, from search engines to academia to even malware development,” Lefebvre said. While “still very much in the early innings on realizing AI’s full potential, the financial industry is poised to be a prime beneficiary of AI.”

The technology behind ChatGPT

ChatGPT pulls information from trillions of data sets across the internet. Users can ask questions or make requests, and the chatbot then uses context-based learning to generate a response.

“What people have been surprised by is seeing that you are talking with [a bot] that has started to think like a human,” Shuki Licht, senior vice president and chief innovation officer at technology provider Finastra, said. “The team in charge of ChatGPT spent lots of time on making the answers correct, but also getting the context based on historical data and relation between words,” he said, adding the quality of intelligence is far superior to what has been seen previously in a chatbot’s ability to solve complex problems.

The engine behind ChatGPT is Generative Pre-trained Transformer 3 (GPT-3), with an updated version of the technology, GPT-4, expected to launch this year. The use of the technology has expanded rapidly, with OpenAI rolling out two sets of paid models for premium services due to server overload within the last month.

Despite the advanced capabilities of the chatbot, its ease of use by the general public has been one of the reasons it has captured audiences in the tech world, Abrar Huq, co-founder and chief revenue officer for AI contract analysis fintech Arteria AI said.

“It’s a very simplistic, but highly effective way in which to provide the general public with access to these models, which may have previously been locked in code on someone’s computer or in a lab somewhere and wasn’t the most accessible,” Huq said.

The race to AI in financial services

ChatGPT is unlocking a world of possibilities in the financial services industry.

“The real question is, what kind of innovations are these generative AI tools likely to lead to?” AI solution Fairplay founder and Chief Executive Kareem Saleh said. What ChatGPT and OpenAI did was captured the imagination of the public through “remarkable demonstration of what’s possible.”

Now, financial institutions that have released AI chatbots, virtual assistants and call center automations need to continue their investment in AI in an industry that often has fallen behind.

“Typically, it’s going to be the most highly innovative, most highly technologically sophisticated [FIs] that are going to be experimenting with these techniques,” Saleh said, but there is no longer time to be last. “The nature of AI is adapt-or-die because the pace of change keeps getting faster.”

AI players

Wells Fargo, for one, is experienced in AI and has a pipeline of AI innovation two to three years ahead of any launches, said Chintan Mehta, chief information officer and head of digital technology and innovation. The $1.8 trillion bank works with Stanford University and the Massachusetts Institute of Technology to ensure potential problems, knowledge and hypotheses are worked out ahead of investment.

“[Less] understood models are going to be problematic in the future, is the basic premise,” Mehta said. “We don’t know what we don’t know.”

Through innovation investments, the bank has the strategy within nonlinear AI experiences, such as getting answers to specific questions through a bank’s chatbot, he told BAN.

A specific nonlinear product is Wells Fargo’s virtual assistant, Fargo, which is powered by Google Cloud, he said. The product, in pilot with its employees, is set to launch in April.

As more industries, including financial services, gear up to offer nonlinear experiences, the timing of the ChatGPT launch “makes perfect sense,” he said. “They had to get [the technology] out there in front of customers so that they can get the feedback, get the reaction, but at the same time, build the reinforcement pipeline around it.”

The following FIs also have launched AI-driven virtual assistants:

What does ChatGPT have to say?

Now, financial institutions don’t have to do all the innovating. In fact, ChatGPT itself shared use cases within the industry including chatbot support and fraud detection.

“Overall, ChatGPT can be a valuable tool for financial services institutions, helping them to streamline operations, improve customer service and make better-informed decisions based on data analysis and insights,” ChatGPT told BAN.

Fintech Fairplay, for example, used the tool to work through a software engineering problem. Listen as Fairplay’s Saleh explains the use case:

The future of AI 

Potential uses for AI technology are surpassed only by its still growing adoption rate via the general public, signaling that ChatGPT may be a player in the AI space for years to come if the technology continues to keep pace with consumer expectations.

“Several years from now, we will likely look back at ChatGPT as the AI watershed moment,” SEI Sphere’s Lefebvre said. “It was initially reported that ChatGPT achieved 1 million users in just five days; fast forward to this month, ChatGPT has apparently amassed 100 million users in just two months.”

A possible use of ChatGPT may come in customer service, as banks see the need for client queries to be answered on nights, weekends or holidays, when it is less likely that customer service reps are working. If OpenAI can develop ChatGPT with voice capabilities, banks may be looking at a 24/7 service, Alex Lawrence, professor at Weber State University in Ogden, Utah, said.

“[AI customer service reps] are always in a good mood, have all the answers that I want, and there are no hold times, which means [there are] no hours of customer service. I can call at 3 p.m. or 3 a.m. and get the information I need,” he said. “On the business side, why wouldn’t I want to have a 24/7 customer service source that doesn’t need benefits, doesn’t call in sick and isn’t rude, and always has the right things that the customer wants?”

When a bank can tend to your needs on New Year’s Day, Thanksgiving, Christmas and other off hours, “that feels like the future,” Saleh added.

Bank Automation Summit US 2023, March 2-3 in Charlotte, is a crucial event on automation technology in banking. Wells Fargo Chief Technology Officer Steve Hagerman will speak at the event. Learn more and register for Bank Automation Summit US 2023.