This consumer is frothing mad and it’s the fault of bad AI. Or bad programming, which inevitably has AI on or about its person.
I ordered something from another country that I couldn’t source at home. This was on a reputable platform but the delivery turned into an unfortunate sequence of events.
My order was received, processed, and arrived in Canada in a couple of days. Awesome efficiency. On Tuesday, I received notification from a certain company that my delivery was scheduled for Wednesday. Twelve minutes later it was rescheduled for Thursday, no reason given. Still, I’m basking in the glow of effortless, international commerce.
…until I realize I have to go to work halfway through the scheduled delivery window. Although there’s a friendly ‘reschedule delivery’ button in the notification, I don’t. The most frequent spam email I get is ‘you have a package’ posing as being from this certain delivery company, so I’m wary of clicking on links in the notice. If this is legit, I’ve got a 50% chance of being there to receive the package.
Of course, I’m not. So I leave a note at the door, asking for the package to be left even though the notification says a signature is required. When I return home , there’s no package.
I must reschedule delivery. I click on the email link, and to my dismay find:
1. I have to join some data-grubbing club/app/membership to reschedule my delivery, and
2. I have to pay to reschedule delivery.
I do both of these things because the alternative is never getting my package. I request delivery the following Tuesday.
Then it gets worse.
Not long after paying for rescheduling, I get a notification that my delivery is rescheduled for Monday. I’m not available on Monday, that’s why I asked for Tuesday. I go through the reschedule routine again, agreeing to pay the fee for rescheduling once more. I feel powerless, fearing I will never see my package.
A confirmation email arrives. My package will be delivered on the Tuesday requested. I begin to believe I have conquered the system. Meanwhile, I’m welcomed to my delivery service account, which I have no use for, or desire to have. I plan to cancel it as soon as I get my package.
Monday rolls around and I get a message that there are duties due on my delivery, but I can pay the driver. I do nothing, planning to pay the driver, as indicated. Now I understand why the previous delivery couldn’t be left at the door. Had I known there was duty due, I wouldn’t have asked for it to be left.
The same email confirms that my delivery will be on Tuesday. I continue to enjoy things working out.
Tuesday dawns a different story. The delivery is rescheduled, for no apparent reason, to the next week.
I need to call someone. Searching through all the previous emails, I find no 1-800 number for inquiries, customer service, or terminally screwed up situations. I find lots of direction to online help, but have had enough of ‘the automated system’.
Google finds me a number to call. It’s answered by an AI.
AI: ‘Please provide your 9 digit tracking number.
My tracking number has 18 digits (including letters), so I say, “I don’t have it.”
AI: ‘It’s right beside the barcode.”
There’s no barcode in my email. So I repeat my 18 digit tracking number to the AI.
The AI repeats it back to me perfectly.
‘Yes,’ I say, ‘that’s it,” feeling we have made a connection.
AI: “That is an invalid tracking number.”
Then the AI has the audacity to suggest this would be simpler if I used their online system.
I lose it and scream at the AI (while feeling somewhat weird yelling at a machine) that the online system is useless and wrong. It has the decency to agree that I need to speak to a human and finds one, who is quite helpful, and recognizes my tracking number.
On the eve of the newly scheduled delivery date, I wonder who will be dominant – the pleasant human I talked to who seemed to have the ability get the schedule right, or the online system that controls the incredible logistics of millions of parcels arriving at millions of locations across the world. It is complex enough to require AI to coordinate. The question is, can AI manage the system in a way that requires the occasional human interaction to get it right?
The story has a happy ending. The parcel arrives on Wednesday, like the human said it would. The driver (also human) tells me there was no previous delivery attempt. This sounds like the ‘online system’ was making arrangements on its own. Could it have gone into an endless loop of rescheduling, for reasons only it understood?
In the end, everything got where it needed to go. Would it have, if the online system and call attendant AI managed the process? It doesn’t seem like it. There may be some logicaly explanation for my experience but from the receiving end, it made no sense.
I made a discretionary purchase, hardly vital. What if the system controlled an emergency call centre, or distribution of vital supplies?
We need to be sure, before we let intelligence systems run our lives, that they are intelligent. Or pull the plug.