Big corporations are gaga over artificial intelligence (“AI”). It is estimated they have invested almost a trillion dollars in the last few years to make it work. They anticipate a 1% annual boost in global labor productivity by 2030, along with $13 trillion in additional economic activity. Also, “By 2030, an estimated 92 million jobs worldwide will be displaced by AI, according to the World Economic Forum’s Future of Jobs Report 2025.” CEO’s believe that AI will be a cash-cow machine that also frees them from dealing with the trials and tribulations of human employees. But what about the person who needs help with a non-routine customer service problem?
AI from a customer perspective. As corporations reduce or eliminate the ability of consumers to access human customer service agents, they will encounter more dissatisfied customers who suffer from maltreatment inflicted by a monotone, computerized voice that offers hope but creates massive frustration and anger.
Let me provide a real-life personal example of attempting to use AI to have my internet restored. Following the recent turbulent storms in the southeast, several telephone poles in my neighborhood were damaged by high winds and lightning. While the high winds broke several poles, I still had internet service since the wires on the ground remained connected to the poles. When the service provider replaced the poles, it had to remove the cables and then reattach them. During the repairs, my wires were somehow disconnected. I immediately reported the interruption in service on the 800 number on my bill. The monotone AI assistant greeted me.
It was my Waiting for Godot Moment. It welcomed me and told me it would resolve my problem. It asked me to identify my problem from a list of issues, none of which directly applied. I selected the problem closest to my needs. First, it instructed me to reboot my router, which did not resolve the issue. Then it applied its troubleshooter technology for about twenty minutes. The troubleshooter was unable to reconnect my service. Every five minutes, I would receive a message: “The test is not finished yet, so please hold on.” After more than half an hour, it provided me with another prompt, informing me that I was next in line to receive primer expediting to “one of the most experienced service agents for priority assistance.” It told me my wait time was 39 minutes. I waited an hour for the “experienced agent” and hung up. I tried again several hours later and went through the same process, which ended in futility an hour later.
A few hours later, I went through the same process, but this time I connected to a technically advanced AI assistant that was very confident (“arrogant”) of its vast amount of knowledge. In a superior tone, it informed me that it could not reconnect my equipment because I did not have electricity. I did have electricity, but no prompt would allow me to explain it. It then provided me with detailed instructions on how to locate my “MOP” box, and reconnect to the electric power supply. (“MOP in Telecom[speak] commonly refers to Method of Procedure, a documented process outlining the steps and protocols for carrying out specific tasks or operations in telecommunications.”) It cautioned me to avoid water in the area where I would work. I assume to prevent electrocution. The AI instructions were very detailed on how to locate specific fiber-optic wires and the source of the lost power.
Wanting to avoid electrocution, I decided not to implement its instructions. Instead, I drove to the provider’s substation to find a human. No one answered the doorbell. Upon returning home, I called the provider several more times. The same AI voice provided the same incoherent, risk-filled information. This time, it gave me the option to hold for a live agent or leave my telephone number to receive a callback at some unspecified future date. I left my number.
Being desperate to gain access to the world, I traveled after dinner to the provider’s retail store to report the problem. I was informed by the clerk greeting me that he could not report the problem since the store only sold cell phones.
The next morning. As it was impossible to get past the AI guard dog, I started driving around the area, looking for one of the provider’s trucks that was doing repairs. I spotted one of its trucks. I parked my car in front of it so it could not move. I got out and asked the driver to report my outage. The driver instructed me to call the 800 number I had been trying to reach. I told him I had called that number many times and received no help. He stated, “That is the only number there is. That is company policy.” I told him my tale of woe and begged him to report the outage. After five minutes of humiliating myself, he said he would report it to his supervisor.
My frustration motivated me to contact the county government division that regulates the cable franchise to some extent. The person I spoke with told me that the provider “laid off all the customer service reps, so AI was my ‘go-to’ source for help.”
When nothing happened by 2 PM, I got back into my car and, luckily, found the same truck and driver who were sitting at a 7-11. I asked him to report the outage again. He stated he did report it. Later in the day, I received a callback from a salesperson who had listened to the voice message I left the day before. The salesperson tried to sell me on upgrading my service. I asked her to report the outage. Several hours later, I received a call from a technical person who indicated the driver reported my outage and that I was on the list for repairs.
The humans deliver the service. Later that day, a repairman came to the house and asked to see the MOP. He had special tools to find the fiber-optic signals. (I assume AI thought I had a set of these tools lying around.) He then went outside to trace my fiber-optic signal from my house to the correct telephone pole and set of lines on the street. He climbed the fifty-foot pole and reconnected my system to its rightful place among a fiber corral hosting many sets of fiber-optic cables.
Customers and corporations are in a war between corporate profits and customer service. While AI may be able to calculate massive amounts of data and regurgitate words from a repair manual, its limited understanding of the real world renders it worthless and potentially even deadly when providing customers with information outside of routine tasks, such as paying bills or ordering new products. What is most concerning about AI is that it is too dumb to know what it does not know. Without humans to manage AI, the world will run amok with risk-filled information that will likely cause information chaos and eventually physical chaos, as no one will know how to distinguish good information from disinformation.
All of the above is about routine consumer assistance. What happens when the government uses AI to regulate society? No one in the nation will know whom to trust, what information they can rely on, or worse, what their government is doing to its citizens. Citizens will not have to worry about George Orwell’s “doublespeak,” the nation will be on “triplespeak.”
William L. Kovacs, author of Devolution of Power: Rolling Back the Federal State to Preserve the Republic. It received five stars from Readers’ Favorite. His previous book, Reform the Kakistocracy, received the 2021 Independent Press Award for Political/Social Change. He served as senior vice president for the U.S. Chamber of Commerce and chief counsel to a congressional committee. He can be contacted at wlk@ReformTheKakistocracy.com
