Illustrating artificial intelligence with a digital brain in place of a head atop a suit.
Artificial intelligence (AI), data mining, expert system software, genetic programming, machine learning, deep learning, neural networks and another modern computer technologies concepts. Brain representing artificial intelligence with printed circuit board (PCB) design.
The state of AI in banking
When you publish an estimate of how a technology will affect an industry — say, that 1.2 million jobs will be lost in banking due to adoption of artificial intelligence software — you get a variety of pointed reactions.

Some will say the number is too high and that AI is an overhyped buzzword like Big Data was a few years ago — soon to be discarded for the Next Big Thing. Others will argue that 1.2 million seems like too small a number. Some insist that AI will only do boring, repetitive tasks, leaving the really interesting stuff for humans. Yet others say that's nonsense, that AI can handle complex, creative or emotional work as well if not better than people can.

Bankers themselves, when surveyed, express contradictory views about AI. They like it, and think it will have a big impact, but so far they are not investing significantly in retraining employees to use it.

No one really knows exactly how much disruption and dislocation AI will cause. But it does seem safe to say that financial companies that find smart uses for AI will have an edge over their peers.

What follows is intel about the attitudes of bank executives and employees toward AI, and where the technology is already making inroads.
Chart: Checks and charity
An AI nudge for charitable donations
Artificial intelligence may help rescue donations from paper's retreat, and give a fintech startup an icebreaker with small banks that haven't embraced automated payments.

The link between check decline and donating is a trend that's already shown up in the U.K., which has shown signs that banks may be attracted to technology that makes donation payments easier.

For example: Santander recently launched a test of contactless payments for charitable giving in the U.K.; and iZettle and SumUp have collaborated on a mobile payment app for charitable giving in the U.K..

One newcomer is Payrailz, which launched in November and plans to launch with banks this year. Automated donations aren't Payrailz's core product, but a feature from the company's mix of a cloud and AI-driven payments dashboard for banks that can be an attraction for users. The app can suggest donations for funds that are "left over" after paying bills based on an AI-driven profile of the user's payments, money management, budget and the bank's CRM system.
Chart: Rise of the chatbots
Consumers driving the conversation with AI chatbots
Mastercard has several projects underway that respond to consumer preferences to use connected devices, which impacts how consumers engage with virtual cards And it also is testing the uses and limitations of voice technology for payment execution and ID in the myriad projects underway to place Mastercard into new connected devices.

According to a study of 3,000 consumers conducted by Mastercard and Mercator, two-thirds of U.S. consumers use either voice assistants or messenger/text chatbots. Out of these, one in five use chat for commerce. Those who prefer voice assistants for shopping prefer its ease and speed over texting. Those who prefer chatbots cite text’s accuracy compared to voice as well as privacy, which is why messenger bots are favored for use in the workplace, according to Mastercard and Mercator.

The challenge is how AI, chat and voice change the more scripted interactions consumers have with e-commerce and online banking sites. In the past, consumers and merchants used the language of payments, credit card companies and merchants. Now consumers are more actively driving the vocabulary.

While there is some pressure on merchants to develop strategies for voice and AI, the current use cases are still very basic. The most common use for voice and text-based agents are informational tasks such as internet searches, merchant inquiries, looking for gas stations, directions or commands to make calls, play music or send texts, according to Mastercard and Mercator. But 21% did say they use voice assistants and/or chatbots to shop, buy, pay bills, bank or send money.
AI slide show chart 1
Bank executives are enthused about AI
Executives say AI will transform their companies and create more jobs. A common refrain among bankers is that AI will remove “grunt work” and free human beings to do more interesting tasks.
AI slide show chart 2
Analysts predict bank jobs will be eliminated by AI
Teller jobs will be hardest hit by the advent of AI, according to researchers at Autonomous. These jobs are already disappearing thanks to the rise in mobile banking and the shutting down of branches. Compliance officers, called "middle-office" workers in this survey because they straddle the outside and back-office worlds, and loan officers, who are included in the back-office category, are among the other job types at risk.
AI slide show chart 4
Executives doubt workers are ready for AI
Bank leaders say their workers are not ready to work with AI. But somewhat counterintuitively, few are planning to invest in helping employees obtain the needed skills.
AI slide show chart 5a
Employees lack confidence about working with AI
Echoing their bosses, only 40% of surveyed bank employees say they are very confident in their skills and abilities to work with intelligent technologies. Millennials are the most upbeat, but even in that age group only 47% say they are very confident.
AI slide show chart 3
Bank employees want to master the game
Still, employees hope AI will bring opportunities for them and think it will make their jobs more interesting. They’re also keen to gain the needed skills to use AI.
AI slide show chart 6
The AI tech that's eliminating jobs
Regardless of how anyone feels about it, AI is already starting to do some of the work of humans in financial services. Chatbots and virtual assistants (such as Bank of America's erica) will do some customer service work. AI-based fraud and regtech software is doing security and compliance tasks (at BBVA, for instance). And it's helping make loan decisions, at Ford Motor Credit, for example.
This article originally appeared in American Banker.