Financial Executives Are Intrigued with 'Artificial Intelligence': Some Envision Automated Teller Machines That Can Talk in Different Languages and Robots That Can Give Advice
Furno, Daryl, American Banker
Some financial services executives are skeptical of the promised benefits of "artificial intelligence" to the banking industry. Others are confused. But nearly all are intrigued at what it could mean.
It appeals to them to coniure up images of automatic teller machines able to converse -- in a choice of languages -- with bank hall patrons. Or images of equally talkative -- and equally robotic -- financial "experts" able to analyze the outcomes of investment decisions for potential customers. Or machines capable of reading and verifying a signature. Or computers that write computer programs.
That such possibilities are of keen interest to executive in the banking/financial services industry is clear to Halbrecht Associates based on an informal telephone survey the firm recently completed. Such interest, of course, comes as no surprise. The banking field has been on the cutting edge of technology for more than two decades, and today it is virtually "inventing the fulture" with respect to its core specialty, telecommunications. But what about "AI"?
We recently spoke to nearly a dozen senior executives at major banking and financial institutions -- as well as to an equal number of leading academics and consultants who, like Halbrecht Associates, work closely with the industry. What emerged from these interviews does not provide answers to all the hard questions of how real, how soon, or how much. Instead, additional questions arose, some of them as knotty as the first. But the range of the questions and concerns do provide considerable insight into the current state of artificial intelligence (AI) in banking. 'Visions of Sugar Plums'
Many of the executives surveyed were skeptical of the benefits of AI. Some had what one professor called "visions of sugar plums dancing in their heads" when they imagined their banks relying totally on AI as in the virtually written about in the press. But the majority of those with whom we spoke were cautious, and they tended to address immediate, practical concerns.
An ececutive vice president for information services with a leading midwest bank noted:
"Why should we bring AI into banking? It would seem so much harder to do than with so many other industries. We have enormous investment in data bases, software, and our present computers. And this is not the right environment to implement AI techniques. They're trying to figure out how to interface AI hardware and AI software with conventional hardware and software. What they really want to do is convert their data bases into 'knowledge bases' that can be used to build AI systems. But the problem is that you cannot do that with current computers, and the current computers are needed to handle large data bases."
Other essentially agree, but for different reasons. According to Alex Jacobson, President of Inference Corporation, a Los Angeles-based AI software company:
"I think that the managerial attitudes of the financial services industry are inimical to the use of new technologies like AI. They tend to throw people at problems, instead of utilizing the best technology available. They are cautious to the point of fearfulness, and there have been a number of cases where industry leaders have backed off on new technology when presented with the need to invest resources to acquire the technology rapidly. They will act only when forced, and with AI they're not in such a position as yet. Right now, the industry is in a riskaverse posture.
One respondent, Scott Abbey of Morgan Stanley, however, simplifies the problem and argues that bank management should not be put off by AI. He believes that management acceptance of AI technology needn't require changes in attitude or management style. Abbey says: "It should only be another tool in their repertoire. It's only another programming language." 'Expert Systems'
Some executives are wary of the scale of investment required, realizing perhaps that only a full-scale research program would be able to produce meaningful results, and that these results would be difficult to predict. …