JPMorgan Chase has rolled out a genetic artificial intelligence assistant to tens of thousands of its employees in recent weeks, the initial phase of a wider plan to roll out the technology across the massive financial giant.
The program, called LLM Suite, is already available to more than 60,000 employees, helping them with tasks such as writing emails and reports. The software is expected to eventually be as ubiquitous at the bank as the Zoom video conferencing program, people with knowledge of the plans told CNBC.
Rather than developing its own AI models, JPMorgan designed the LLM Suite to be a gateway that allows users to use external models of large languages—the complex programs that power its AI generation tools—and released it with its LLM ChatGPT maker OpenAI, people said.
“Ultimately, we’d like to be able to move pretty smoothly through the models depending on the use cases.” Theresa HeitzenretherJPMorgan’s head of data and analytics, said in an interview. “The design must not belong to any model provider.”
Teresa Heitsenrether is the company’s head of data and analytics.
Courtesy: Joe Vericker | PhotoBureau
The move by JPMorgan, the largest US bank by assets, shows how quickly genetic AI has swept US companies since ChatGPT’s arrival in late 2022. Rival bank Morgan Stanley has already released a pair of OpenAI-powered tools for its financial advisors. And consumer tech giant Apple he said in June that he was embodiment OpenAI uses models in the operating system of its hundreds of millions of consumer devices, greatly expanding its reach.
The technology — hailed by some as “Cognitive Revolution“In which tasks formerly performed by knowledge workers become automated — they could be as important as the advent of electricity, the printing press and the Internet,” JPMorgan CEO Jamie Dimon said in April.
It will likely “increase almost every job” at the bank, Dimon said. JPMorgan had about 313,000 employees in June.
Ban ChatGPT
The bank is giving employees what is essentially OpenAI’s ChatGPT in a JPMorgan-approved wrapper more than a year after confined employees from using ChatGPT. That’s because JPMorgan didn’t want to expose its data to outside providers, Heitsenrether said.
“Since our data is a key differentiator, we don’t want it to be used to train the model,” he said. “We’ve implemented it in a way that we can leverage the model while keeping our data protected.”
The bank has rolled out the LLM Suite broadly across the firm, with teams using it in JPMorgan’s consumer division, investment bank and asset and wealth management, the people said. It can help employees write, summarize long documents, solve problems using Excel, and generate ideas.
But getting it on employees’ desktops is just the first step, according to Heitsenrether, who was promoted in 2023 to head the bank’s red-hot technology adoption.
“You have to teach people how to do it direct engineering that’s relevant to their field to show them what it can really do,” Heitsenrether said. “The more people get deep into it and unlock what’s good and what’s not, the more we start to see ideas blossom really.”
The bank’s engineers can also use the LLM Suite to integrate functionality from external AI models directly into their programs, he said.
‘Exponentially Bigger’
JPMorgan has been working on traditional AI and machine learning for more than a decade, but the arrival of ChatGPT forced it to pivot.
Traditional or narrow AI performs specific tasks that involve pattern recognition, such as making predictions based on historical data. However, Generative AI is more advanced and trains models on huge data sets with the goal of generating patterns, i.e. how human-sounding text or realistic images are formed.
The number of uses for genetic AI is “exponentially larger” than previous technology because of how versatile LLMs are, Heitsenrether said.
The bank is testing several instances of both forms of AI and has already put some into production.
JPMorgan uses genetic AI to create marketing content for social media channels, map itineraries for its clients travel agency acquired 2022 and summarize meetings for financial advisors, he said.
The consumer bank is using artificial intelligence to determine where to put new branches and ATMs, taking satellite imagery and into call centers to help service staff find answers quickly, Heitsenrether said.
In the company’s global payments business, which moves more than $8 trillion around the world every day, artificial intelligence helps prevent hundreds of millions of dollars in fraud, he said.
But the bank is more cautious about genetic AI that directly touches the individual customer because of the risk of a chatbot giving bad information, Heitsenrether said.
Ultimately, the AI manufacturing field may evolve into “five or six big fundamental models” that dominate the market, he said.
The bank is trying LLMs from US tech giants as well as open-source models to integrate into its portal next, said the people, who declined to be identified talking about the bank’s AI strategy.
Friend or Foe?
Heitsenrether laid out three stages for the evolution of genetic AI at JPMorgan.
The first is simply making the models available to workers. The second involves adding JPMorgan’s proprietary data to help boost employee productivity, which is the stage that has just begun at the company.
The third is a bigger leap that will unlock much greater productivity gains, which happens when genetic AI is strong enough to function as a standalone agents that perform complex multi-step tasks. This would make employees more like managers with AI assistants at their command.
Technology will likely empower some workers while displacing others, changing the makeup of the industry in ways that are hard to predict.
Banking is the most susceptible to automation of all industries, including technology, healthcare and retail, according to the consultancy Accenture. AI could boost industry profits by $170 billion in just four years Citigroup analysts said.
People should think of genetic AI “as an assistant that takes away the most mundane things that we all wish we didn’t have to do, where it can just give you the answer without digging through spreadsheets,” Heitsenrether said.
“You can focus on higher-value work,” he said.
— CNBC’s Leslie Picker contributed to this report.
