Today we are seeing an increasing amount of news and commentary about the necessity for artificial intelligence (AI) to intervene in or perhaps even replace live human interaction in areas such as sales and customer service. I beg to differ with such assessments–but not for reasons of “man versus machine” or other such cliches. My reasoning is actually very practical.
If you think about it, yours might be, too. Recall the last time you encountered an AI-powered customer support solution. “Please say your account number.” “What kind of support do you wish?” “Do you wish to find out about our premium offers?” and so on. The usual end product of such a call–at least for me, and possibly for you, too–is angrily screaming into the phone, “operator!” or “agent!”
The individuals and organizations advancing these goals for AI say that technology is reaching a level where it can perform these functions better than humans, and thereby eliminate human error. Additionally in today’s digital world there is simply too much data for people to confront, organize and analyze, and artificial Intelligence is the answer.
There are a number of high-profile AI projects out there. IBM Watson is being utilized in customer service, music delivery, hospitality, and even healthcare. Salesforce Einstein is a collection of AI services designed to analyze data and provide direction for customer service. Chatbots are designed to convincingly simulate how a human would behave in a conversation, either through text or speech. There are other examples.
Ask anyone who knows me–I’m the first person who will pick up and use technology wherever I can. I opened one of the very first Apple dealerships in Vienna, Austria in the mid-1980s. I was one of the first people I know to own a smartphone. I implemented long-distance online conferencing in my company years ago, as soon as it was available. I myself am a software developer and have created several innovative products still selling well today.
But I strongly believe that we should never replace human interaction. Despite my adoption of technology, you’ll always find me seeking out a human for tech support, or any kind of customer service.
My reasons are simple: No matter how sophisticated the machine or the programming, any answer or response it provides is based strictly on 2-valued logic: “yes” or “no.” An answer is “wrong” or “right.”
A simple example would be an elderly man coming to a bank over the weekend, needing cash. He’s been a good customer of the bank for 20 years–they know him very well. But when he comes up to the ATM to obtain money from his account, if he doesn’t know his PIN, he’s not going to get money, period. Whereas if he came in during banking hours, he wouldn’t even need an ATM card. A teller, recognizing him, would simply have him fill out a withdrawal slip and would provide the needed money, no questions asked.
This kind of thing could certainly happen in a business setting, too. Let’s say that your company’s database has become corrupted and you cannot find the reason. One of your IT staff is tasked with reaching out to the database vendor for tech support, for which your company pays. The call is answered by an AI-powered software solution. When prompted for your company’s account number, your IT staffer provides it. The robot-voice says, “I’m sorry, that is an invalid account number. Please say it again.” Your IT staff person does–louder and clearer. Once again, it is rejected.
It takes about an hour for the IT staff person to finally reach a human. When she does, she finds out that your company’s account number had been entered into the vendor’s system incorrectly. Because of the “yes” or “no” nature of automation, no understanding was possible. A human, however, was able to sort this out in less than a minute.
Which leads me to my other major reason I don’t believe artificial intelligence should completely replace humans: only humans are capable of understanding. A machine can only act upon the data and options programmed into it, and is totally incapable of understanding and reason.
A human can pause, reflect, think, and decide how to respond to another person–be it a customer, a prospect, or simply someone needing help.
Perhaps someday AI will be capable of such things. But this isn’t on the horizon in the foreseeable future.
Relating to a False Model
An AI solution can only respond to someone based on who it thinks it’s talking to. A sophisticated solution is going to be programmed to respond to some kind of average individual–for example, a buyer–that would reach out to a company. We could call this programmed individual a model, and in many cases it’s not going to be the right model.
A fantastic example of such a model is homo economicus. Homo economicus is a model evolved by economists, originally in the 19th century, to describe the economic behavior of the average person within an economic framework. Many major schools of economic theory utilize homo economicus within their calculations, analyses and predictions.
The precise opposite of the homo economicus model is the one utilized in the Austrian School of Economic Thought: the free-will individual. In the Austrian School, the individual acts as they do in real life: unpredictably. A human can make any decision they like–and can never be totally and accurately predicted, especially under one encompassing model.
I myself am a great example of a total exception to the homo economicus norm. I may be the only Austrian who, at over 50 years of age, gave up just about everything at home to move to the U.S. and start a business, practically from scratch. The average person of my age back home in Vienna would be looking forward to retirement–not throwing it all in to move to a different country.
How Should AI Be Put to Use?
So if I am saying that AI shouldn’t replace human interaction, how should it be used?
Simple: I believe the amazing power of AI should be utilized to help people to make the right decisions. Let us certainly utilize it to help analyze the terabytes of data out there, so that we can make informed decisions about commerce, sales, marketing and other important fields.
In other words, we can take advice from artificial intelligence–but it can never replace us.
Well beyond the many possible philosophical reasons for this standpoint, the practical one is that only humans can understand, reason and truly help. It is humans that should remain in sales and customer service–for the sake of customers and for companies.