Financial services shun AI over job and regulatory concerns

European fintech executives say financial services are not effectively implementing artificial intelligence, even though there is growing evidence that the popular technology will boost productivity and lower costs.

Concerns about job losses, regulatory issues and institutional inertia are some of the factors deterring bankers from fully implementing the systems that underpin products like ChatGPT.

“The big banks are definitely not going to adopt (the technology) as quickly as any of the fintechs,” said Tom Blomfield, co-founder of Monzo and a group partner at Silicon Valley startup incubator Y Combinator. But generative AI “will make banks more efficient and able to deliver the same products at a lower cost.”

Only 6 percent retail banks are ready to implement AI on a large scale across their business, according to a Capgemini study. However, McKinsey estimates that it could add up to $340 billion in value annually to the global banking sector, equivalent to about 4.7%. total industry revenues.

Many say the technology, which can answer questions and analyze vast amounts of text and numeric data in seconds, could cut costs across the industry, but there are concerns that disruption will lead to job losses.

“People don’t understand that this is a productivity tool,” said Nasir Zubairi, CEO of fintech accelerator Luxembourg House of Financial Technology. “They still genuinely believe that it will put them out of work.”

He added: “Traditional banks are analog by design, and the transition from analog to digital has always been difficult.”

Zubairi, speaking at the Financial Times’ TNW technology conference this month, gave the example of money laundering controls, where institutions typically hire staff to review spreadsheets looking for unusual activity.

He said that when he demonstrated to one institution how this could be improved with a bespoke AI model, which he estimated could immediately save up to “€450,000 a year in salary”, the proposal was rejected.

“People don’t like to fire people,” he added. “They want to protect the function of their job, and if they have to fire the people on their team who do that work, they’re potentially at risk as well, because the leadership or their authority is also eroded in some way.”

According to the Bank for International Settlements, central banks have recently been urged to “up their game” when it comes to artificial intelligence, as the technology can provide productivity gains but also carries risks such as misinformation and vulnerability to hacking.

A common problem with large language models, the technology that underlies most generative AI products, is their tendency to “hallucinate” and take inaccuracies as fact. They are also known to generate information based on data they have been trained on, leading to concerns about sensitive or secure information.

“You don’t necessarily have to dismiss AI, but there is hesitation,” said Wincie Wong, director of digital solutions at NatWest, who called for an assessment of the technology’s risks, ethics and vulnerabilities before it is deployed. “Ultimately, we are one of the big banks and many customers keep their data and finances safe with us. We have to respect that.”

Customer service is one of the areas most disrupted by AI tools that can talk human-like and respond to queries. For more than a decade, digital banks have used machine learning to sort online questions, often directing customers to a live customer service agent.

However, LLM-based bots can understand a wider range of queries, regardless of how they are phrased, and can make decisions such as ordering a bank card, eliminating the need for human intervention.

“I really think the vast majority of customer service roles will disappear in the ‘next 12 months to five years,’” Monzo’s Blomfield said.

Many banks and fintech companies, including Klarna and NatWest, are already using AI chatbots for customer service. NatWest’s Wong said they have made huge strides in generative AI with their AI Cora service, receiving more than 11 million chats in a year, more than half of which did not require human intervention. In 2017, the service was receiving 1,000 chats per month and required intervention.

Swedish fintech Klarna said its AI assistant can do the work of 700 customer service employees and resolve queries in less than two minutes, up from 11 minutes previously. As a result, the company expected to save $40 million in customer service costs this year.

But Wong said training the models on nuances was key to the project’s success. For example, understanding that a change of address could have emotional implications, such as a family bereavement.

“Understanding the psychology behind it was really important, and if you don’t get it right, you can, to put it bluntly, piss off a lot of customers,” she added.

Banks also had to be careful when adopting new technology while adhering to stringent industry rules and navigating an unfamiliar regulatory environment.

In a landmark 2022 ruling, a Dutch court ruled in favor of neobank Bunq after it sued the Dutch central bank for banning it from using artificial intelligence to conduct money laundering checks.

Last month, regulators lifted restrictions on German fintech N26 after it improved its controls. For years, the bank had been restricted from signing up new customers because of weak money-laundering controls and faced millions of euros in fines for persistently late filings of suspicious activity reports.

Carina Kozole, chief risk officer at N26, said the company worked closely with regulators to create an artificial intelligence model that would assess whether a new customer was a criminal. As a result, the number of such cases on the platform decreased by 90 percent.

“If we don’t apply AI to the industry, we won’t be here in a few years,” she added. “We need to show the benefits and how we can ensure compliance if we use AI.”