The ChatGPT hype has sparked an Artificial Intelligence (AI) arms race among big tech companies such as Meta, Google, Apple, and Microsoft. Microsoft has been aggressively positioning itself in this space with a stake in OpenAI, the creator of ChatGPT, with a multi-year, multi billion investment plan. This intensified focus on AI and its utilization is not reserved to big tech companies or the technology sector. Major financial institutions are zeroing in on advanced AI products and services, too.
JPMorgan (JPM) CEO Jamie Dimon recently acknowledged the benefits of AI during the bank’s last annual meeting, calling the technology “extraordinary and groundbreaking,” further indicating that “AI has helped us to significantly decrease risk in our retail business and improve trading optimization and portfolio construction.”
JPMorgan is currently working on more than 300 AI use cases and has spent over $2 billion building cloud-based data centers to modernize a significant portion of its applications to run in both their public and private cloud environments. “AI and the raw material that feeds it, data, will be critical to our company’s future success,” stressed Dimon.
Following Dimon’s sentiment on AI, the news that JPMorgan applied to trademark a product called IndexGPT and is developing a ChatGPT-like software service that leans on a disruptive form of artificial intelligence to select investments for customers, comes as no surprise. The fact that JPMorgan, the biggest U.S. bank by assets, is leading an undertaking in Generative AI and Generative Pre-trained Transformer (the “GPT” in Chat GPT) is quite telling, but it is not the only financial institution trying to capitalize in the advancements in AI.
Other examples include:
- Bank One Zero developing a Generative AI chatbot for customers’ interactions
- Kassisto launching KAI-GPT, a banking-specific Large Language Model (LLM)
- Bloomberg releasing a research paper detailing the development of BloombergGPT, a new large scale generative AI model
The use of Machine Learning (ML) and AI models in the financial industry is not new. Financial institutions have acknowledged the power of this technology and have been implementing it in various services and products for decades. Granted, these models may look primitive compared to the capabilities of AI due to the advancements in computing power, LLMs, and GPTs.
The current state of AI-powered applications in the financial industry
Financial institutions have been utilizing AI and automation across a variety of services and products – from risk management to portfolio optimization to credit scores and customer service. Robo advisors and customer assistance chatbots are probably the most common ones:
Betterment was the first robo advisor launched in 2008, with the initial purpose of rebalancing assets within target date funds to help manage passive, buy-and-hold investments through a simple online interface.
The technology behind Betterment was nothing new. Human wealth managers had been using automated portfolio allocation software since the early 2000s. But until Betterment launched, they were the only ones who could buy the technology, so clients had to employ a financial advisor to benefit from that innovation. Today, most robo advisors use passive indexing strategies that are optimized using some variant of modern portfolio theory.
Robo advisors have not replaced human advisors. Wealth management firms, including Morgan Stanley and Merrill, offer simple robo advisor services, but that hasn’t stopped their human advisors from gathering billions of dollars more in assets.
Chatbots and customer assistance
Among the top ten commercial banks in the U.S., all use chatbots of varying complexity to engage with customers. Much of the industry uses simple rules-based chatbots with either decision tree logic or databases of keywords or emojis that trigger preset, limited responses or route customers to Frequently Asked Questions (FAQs). Other institutions have built their own chatbots by training algorithms with real customer conversations and chat logs, like Capital One’s Eno and Bank of America’s (BoA) Erica.
Launched in 2018, Erica has become one of the most-accessed virtual banking assistants, helping 32 million customers with over 1 billion interactions. In early 2023, BoA enhanced the chatbot to give customers even more personalization and tailored product recommendations. Users who begin interacting with Erica online can switch to speaking with a human agent when they need more help.
The banking industry’s digital transformation, accelerated by the pandemic, have changed the way many consumers manage their financial lives. BoA recognized that it’s not just younger generations who are using Erica; older clients are also taking advantage of the technology. But the bank realized that consumers of all ages still crave some human interaction. The KeyBank 2020 Financial Resiliency Survey indicates that Millennials and Gen Zers prefer a combination of digital and in-person banking more often than older Americans.
In addition to BoA, other major banks like Wells Fargo and Truist have recently introduced AI-powered virtual assistants. AI, though, is unlikely to replace skilled human workers. Computational algorithms lack common sense and empathy, so humans still have a role in customer relationships. Employing the latest advancements in GPT, financial institutions are currently developing applications which will take these services to the next level, making them more powerful, efficient, and effective.
The future of AI-powered services in the financial industry
A recent study from the Capgemini Research Institute, titled “Why consumers love generative AI,” supports financial institutions’ interests in developing Generative AI services. The study found that adoption and awareness of the technology is relatively high among all age groups and geographies, with more than half (51%) of respondents saying they are aware of the latest trends and have explored the tools. More surprisingly, the study found that consumers have high trust levels in financial advice dispensed by generative AI platforms: 53 percent of consumers trust generative AI-assisted financial planning.
The next generation of robo advisors
Financial advisors have long feared the arrival of technology good enough to displace their role in markets. Those fears have largely yet to materialize. JPMorgan, which employs 1,500 data scientists and machine-learning engineers, is testing “a number of use cases” for GPT technology. “We couldn’t discuss AI without mentioning GPT and large language models,” said Lori Beer, global tech chief, “We’ve recognized the power and opportunity of these tools and are committed to exploring all the ways they can deliver value for the firm.”
According to the trademark filing, IndexGPT will tap “cloud computing software using artificial intelligence” for “analyzing and selecting securities tailored to customer needs.”
Bloomberg has been working on a new AI model that aims to revolutionize the finance industry, called BloombergGPT, a new LLM that has been trained on a massive amount of financial data to assist with a variety of natural language processing (NLP) tasks within the financial industry. BloombergGPT is an advanced machine learning software that can rapidly analyze financial data to assist with making risk assessments, judge financial sentiment, and potentially even automate accounting and auditing tasks and more.
The complexity and unique terminology of the financial industry requires an AI that is specifically trained with financial datasets. BloombergGPT will unlock new opportunities for amassing the vast quantities of data available on the Bloomberg Terminal to better help the firm’s customers, while bringing the full potential of AI to the financial domain. The BloombergGPT model outperforms existing open models of a similar size on financial tasks by large margins, while still performing on par or better on general NLP benchmarks.
Table 1: How BloombergGPT performs across two broad categories of NLP tasks: finance-specific and general-purpose (excerpts from research paper.)
Robo advisors may not have put financial advisors out of business, but these GPT-powered applications might.
The next generation of chatbots and customer assistance
Conversational AI chatbot developer, Kasisto, has recently, launched KAI-GPT, the world’s first banking-specific LLM, which empowers banks with the potential of generative AI to provide human-like, financially literate interactions at speed and scale. The firm’s first application to run on KAI-GPT is KAI Answers, a conversational response system that helps bankers locate, interpret, and understand the information from a variety of sources, including policies, regulatory filings, procedures, web content, and complex financial products. Westpac, the Australian banking giant that is both an investor and client of Kasisto, is already implementing KAI Answers.
One Zero bank has initiated a pilot program involving 450 customers, with plans to roll out the service to the public in Q4 2023. It will be one of the first to utilize GPT chat platforms for private banking and will also give money management advice. The GenAI-based assistant will provide responses encompassing a wide range of inquiries including simple questions such as available account balance, accrued deposit interest, upcoming credit card payment, as well as more complex queries like comparing expenses between specific months of different years. Additionally, the system will utilize predictive capabilities, leveraging AI technology developed by the bank, to address predictive questions. During the pilot phase, the bank will focus on training the model specifically to address banking-related questions while programming the machine to avoid providing responses that may have legal or ethical implications.
Challenges and risks
The Consumer Financial Protection Bureau (CFPB) is monitoring banks’ increasing use of AI-powered chatbots amid a surge of complaints from frustrated customers trying to receive straightforward answers from their financial institutions or raise a concern or dispute. The CFPB has raised a few concerns:
- Noncompliance with federal consumer financial protection laws
- Diminished customer service and trust
- Harm to consumers: When chatbots provide inaccurate information regarding a consumer financial product or service, there is potential to cause considerable harm
As Jamie Dimon said, AI is “extraordinary and groundbreaking,” and can revolutionize financial services and products. If we are mindful of the challenges and risks and implement the necessary measures to mitigate them, we can reap great benefits from utilizing AI and GPT.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.