The proliferation of generative A.I. in the past several months has led to an internal disagreement at Eurobank about the future of the disruptive tech.
“Our bank is divided,” said Andreas Athanassopoulos, deputy CEO at the Greek-based bank. “We are like a religious group. We have those who are emphatically in favor, and those who are afraid.”
Business leaders that joined together for a virtual conversation led by Fortune CEO Alan Murray on Wednesday agree that A.I. is an emerging tech that could rival the momentous debut of the iPhone or even the Internet itself. Nearly three-fourths of companies are prioritizing A.I. over all other digital investments, and up to 40% of the workforce could be impacted by generative A.I., Accenture research has shown.
But the same executives struck a cautious tone, especially as it relates to regulation and privacy for large industries like tech and health care, the protection of intellectual property, and private corporate or customer data.
“You can’t depend on regulation and government yet. It is not happening as fast as needed,” said Joe Depa, Accenture global managing director and data-led transformation lead. “It is really important you have a responsible A.I. framework in place.”
Athanassopoulos said deploying A.I. in areas of the financial industry’s business that have high regulatory considerations can be especially difficult. Kyle Corcoran, an investment banker at Morgan Stanley, expressed a similar note of vigilance. “In the nearer term, banking usage is going to be more limited,” Corcoran said.
With an undergraduate degree in computer science, Corcoran is personally interested in the potential for A.I. in coding. But he said what’s being generated requires a human review because “there are hallucinations; there are things that may come into your code that are completely outlandish.”
Another major concern is that if client data is added to A.I., it essentially becomes public and, at worst, could even become the property of the organizations that run those generative A.I. models. That could put reputations at risk—for both clients and firms like Morgan Stanley.
“Trust is paramount in enterprise A.I.,” said Rahul Auradkar, an executive vice president and general manager at Salesforce. To address those concerns with customers, Salesforce opens conversations about A.I. by addressing the principles and guidelines that have been publicly published in order to outline how they responsibly develop and implement generative A.I. Salesforce also works with regulatory bodies to set security standards.
There’s also the issue of C-suite executives and board members pushing for A.I. adoption but not being sure of where to start or how to use the technology effectively. “It is very easy to fall in the trap and solve a small problem with generative A.I. where it is unwarranted and becomes extremely expensive,” said Alexander Statnikov, a general manager at payment processor Square.
“Everybody is starting to experiment now. We’ve gotten past the initial awareness phase,” said Depa. “We are also guiding our clients not to get too caught up in the hype. Gen A.I. is just one example and one piece of many things you have to get right across your data foundation.”
Some leaders during the virtual conversation were more bullish about the potential of generative A.I. Omri Morgenshtern, CEO of Booking Holdings’ Agoda, said his company uses generative A.I. for internal discussions and problem solving. “It brings a lot of context that is very hard to bring very, very quickly, into ongoing discussions,” Morgenshtern said.
Adtalem, which educates nearly 90,000 health care workers, sees great potential for A.I. in marketing and when configuring classes to ensure that the faculty can have the least amount of teaching hours with the most students. That said, “it is really important to ground ourselves in the fact that ChatGPT is not the first large language model that the board has seen,” said Lisa Wardell, a board member at Adtalem.
Across industries, Depa said Accenture sees massive disruption for content creation and the call center function, but A.I. will be harder to use as a tool for fields like health care and software development. That said, there are some encouraging signs that there is an appetite for adoption. One recent study found health care professionals vastly preferred medical advice from ChatGPT over human doctors.
“In a few years from now, this can change how society functions and works,” said Depa. “But in order to do that, we have to resolve some of these cybersecurity issues.”