Laurent - stock.adobe.com
Advanced AI in financial services boosts fraud detection, efficiency
Financial firms plan to invest more into R&D on AI and plan to deploy advanced AI, like deep learning, within the next two years, according to a new survey.
Financial institutions have long used automation and simple AI to help file paperwork and detect fraud. Now, however, more financial services firms use AI for complex cases and have plans to increase spending to develop new AI products and technologies.
Importance of AI in financial services
A recent survey from the World Economic Forum and the University of Cambridge Judge Business School, with support from Invesco and Ernst & Young (EY) indicates that the use of AI by financial institutions is beginning to be widespread.
According to the survey, which included 151 firms, both fintechs and incumbent financial institutions, 85% of respondents currently use AI. Fintechs, compared to incumbents, are slightly ahead in AI adoption. Fintechs are technology-focused and can be startups. Incumbents, meanwhile, are traditional financial institutions.
"I was surprised by how many of the respondents believe that AI is so critical" to their business, said Nicola Morini Bianzino, global chief client technology officer at EY.
About 77% of respondents see a very high importance for AI in financial services within the next two years, and many respondents indicated they will expand their use of AI in risk management, client acquisition, customer service and to generate new revenue through new products and processes.
Despite the importance of AI, the technology is not a large part of total R&D expenditure for most respondents. About 40% of them said they ingest more than 10% of overall R&D resources in AI, although those numbers are expected to rise in the next two years.
Partly, the low percentages are because "there are very, very few organizations on the planet that can do pure R&D," Bianzino said. Those that can't invest in creating new AI technologies or products are focusing on developing new ways to apply existing technologies.
Simple machine learning and automation
Currently, financial institutions largely use simple machine learning algorithms and chatbots, as opposed to complex AI systems that involve deep learning. Within two years, however, many respondents plan to implement NLP and computer vision.
Financial institutions are moving AI from the back end to the front and are beginning to deploy more advanced technologies, Bianzino said.
Still, according to Karen Reichle, vice president of global customer success engagement at RPA vendor Nintex, many financial institutions also use automation and RPA for various purposes. A popular use, she said, is to automate authentication and compliance processes to ensure institutions are following financial regulations.
Political changes typically bring modifications to banking regulations, Reichle said. Using RPA, banks can automatically file paperwork in a way that adheres to regulations. Compared to manually filing paperwork, using RPA helps keep mistakes down and ensures the correct forms are filed, and filed correctly.
Other common uses include AI-enabled data analytics, fraud and anomaly detection, and AI-enabled customer communication channels, according to the survey.
Yet, even as accustomed as financial institutions are to regulatory changes, many respondents in the joint survey said AI regulation would hamper their implementation of the technology. Around 40% said regulation would impede on their deployment of AI, while 30% welcomed the regulations.
"On one end, it will be very welcome," Bianzino said. Regulations around explainable AI, for example, is important, as more explainable systems will give business leaders greater visibility into what their AI systems are doing and why.
Yet, regulations around AI bias and explainable AI will be difficult, if not impossible, to enforce, he added. That's why there is such a split in opinion, Bianzino said.
Nicola Morini BianzinoGlobal chief client technology officer, EY
The survey also found that fintechs expect AI will actually expand their workforce by 19% by 2030. According to Bianzino, AI doesn't necessarily take jobs away. Instead, it changes people's tasks and their job titles, he said.
Still, incumbent financial institutions expect AI to replace 9% of all jobs within their organizations by 2030. Based on the numbers, the survey found that there would be a net reduction of around 336,000 jobs for incumbents, while only an increase of 37,700 jobs in fintechs.
For Bianzino, though, the survey generally points to one thing: "It's very clear AI is not a side show anymore."
Finance industry giants disclose AI challenges